AI Isn’t the Strategy: Fern Potter on Intelligent Assistance, Human Judgment, and the Future of Work
Artificial intelligence may be the most transformative technology of our generation—but according to Fern Potter, AI alone is not a strategy.
In this episode of Signal & Noise, hosts Rio Longacre and Brett House sit down with Fern Potter, Co-Founder of Intelligent Assistance, to explore why the greatest opportunity in AI isn’t necessarily replacing people—it’s amplifying human judgment.
After more than two decades leading strategy, product, partnerships, and commercial growth across agencies, media, and ad tech—including serving as Chief Strategy & Growth Officer at Multilocal—Fern made the leap to entrepreneurship. Alongside her co-founders, she launched Intelligent Assistance around a simple but powerful philosophy: AI creates the most value when it enhances human expertise, context, creativity, and accountability rather than attempting to eliminate them.
The conversation begins with Fern’s journey from agency leadership to founding an AI company at a moment when enterprises are rushing to deploy generative AI. She explains why so many organizations start with technology instead of business problems—and why that approach almost always leads to disappointing results.
From there, the discussion explores what “intelligent assistance” actually means in practice. Fern explains how organizations should determine which work should be automated, which decisions should remain firmly human, and how AI can become a force multiplier instead of another disconnected productivity tool.
Rio and Brett also dive into one of the episode’s central themes: the difference between intelligence and autonomy. Just because AI can make a decision doesn’t mean it should. Fern discusses the critical role of context, accountability, governance, and human oversight as organizations increasingly rely on AI-assisted workflows.
Drawing on her experience helping reshape programmatic advertising through curation and supply-side innovation, Fern shares lessons that extend far beyond media. The trio explores how unchecked automation created inefficiencies and opacity in advertising—and why enterprise AI risks repeating many of the same mistakes if organizations optimize solely for automation instead of outcomes.
The conversation also tackles larger questions about the future of work. What happens to agencies, consultancies, and professional services when small AI-enabled teams can accomplish what once required dozens of people? Which human capabilities become more valuable as technical execution becomes increasingly automated? And how should leaders redesign organizations around “thinking power” rather than simply reducing headcount?
Whether you’re leading AI initiatives, building products, transforming marketing organizations, or simply trying to understand what responsible AI adoption looks like, this episode offers a thoughtful, practical framework for moving beyond the hype toward meaningful business impact.
In this episode, you’ll learn:
Why AI is not a business strategy
The difference between automation, autonomy, and intelligent assistance
Why most enterprise AI initiatives fail to deliver commercial value
How to combine AI with human judgment for better outcomes
Lessons enterprise AI can learn from programmatic advertising
Why organizational redesign matters more than technology deployment
Which uniquely human skills become more valuable in the AI era
How leaders should think about governance, accountability, and trust
If you enjoy conversations about AI strategy, marketing transformation, organizational design, and the future of work, be sure to subscribe to Signal & Noise for weekly conversations with the leaders shaping the future of business and technology.
#SignalAndNoise #ArtificialIntelligence #AI #GenerativeAI #FutureOfWork #HumanCenteredAI #Leadership #BusinessTransformation #Marketing #MarTech #AdTech #DigitalTransformation #EnterpriseAI #Innovation #Technology #IntelligentAssistance
Read the full transcript below.
Brett (00:01.011)
Hey everybody, this is Brett House, Signal and Noise, joined by my co-ho co-host Rio Longanchor. And today's guest is Fern Potter, one of my favorite names in the industry. I mean, you know, Fern Potter. Hard to forget that. yeah, and she lives in the Shires of England. so I mean Fern Potter and the Shires, it's just a good combination. And we met she's the co-founder of a company called Intelligent Assistance.
Rio (00:07.09)
Yeah, they're
Rio (00:12.998)
It's a super cool name, yeah. It's a badass name.
Brett (00:28.927)
which is a new business that combines human and artificial intelligence to help companies solve practical commercial challenges. I've certainly talked to a lot of companies that are trying to do that. it's a critical challenge and opportunity, I think, in our ecosystem and our world right now is how do companies become AI first without drowning in the in the complexity of all this. And it sounds like your organization is trying to help companies do that.
and you founded this with some a couple of others, right? So I was I I did ask if you were the founder or the co-founder, Sarah DeMartin and James Harris. maybe you could introduce and let the audience know a little bit about them. But you've kind of founded it on a pretty deceptively simple premise, right? Is that AI works best when it's a when it's assists people rather than pretending to replace human understanding, right? Human comprehension, that sort of thing, which I thought was just a good interesting.
way of thinking about it and an important way of thinking about it. and and before founding Intelligent Assistant, you built a career across media technology, product and commercial growth. You were at you were the chief strategy and growth officer at Multi Local. So I'd love to hear a little bit about that. It was yeah. And they're a technology led curation intelligence company. So you'll have to define that for the audience, which I think is is certainly interesting. And you've been a prominent voice on programmatic curation, supply side decisioning, sustainability.
Rio (01:38.386)
Yeah, a long time, yeah.
Brett (01:54.019)
and returning more control to media owners, which is something that that someone that we have in common, David Nuremberg, where we met in Miami at the possible event this year at his dinner.
Rio (02:05.701)
That dinner, yeah, the d David's dinner, that was super fun. Yeah.
Fern Potter (02:06.781)
Yeah, we were sat next to each other. It was great.
Brett (02:09.166)
Yeah, and and he's one of those out you know, you kind of provocateurs out there talking about l you know, we need transparency and honesty in in kind of a a system that oftentimes isn't either of those. So welcome to the show, Fern. it's been it's been good getting to know you over the last couple of months and we had a great time in in Possible, and and thrilled to have you.
Fern Potter (02:31.438)
Thank you. Thanks for having me. It's really exciting to meet again and have this conversation. think, as you said, we founded Intelligent Assistance on the premise that there's so much out there, there's so much complexity in the market, but human intelligence is really fundamental to competitive advantage and how businesses grow. And AI is this, and there's a play on words there. We talk about the fact that AI needs IA, right? That it's a requirement to...
have this context, know what good looks like, know what the outputs need to be, and use AI to accelerate that, not necessarily replace it. You talked a lot of, you know.
Brett (03:06.576)
Yeah.
Fern Potter (03:08.666)
and you see this in market a lot, there's so much around efficiencies, headcount costs and cuts, etc. and a huge hyperbole around it. I think when we started this business, it was really to go, look, there's other ways of using it. Yeah, that's definitely happening. You can't deny it. But there's amazing amounts of work out there, tools and solutions that we can bring together in what we call engines to help businesses grow.
and adopt at the rate of which they're ready to do so. I think there's a lot we can discuss around the technical side of things, but also the psychological side of how... my gosh, you're so right. know, as much...
Brett (03:42.353)
Yeah, which is equally which is half the battle, right? Yeah, and and and what in a lot of the clients that I talk to and and their clients as well, you know, struggle with with being able to sort of translate some of the hype to to reality on the ground within these organizations. And oftentimes companies are just well behind the curve and they don't know necessarily where to start.
Rio (03:59.216)
Was a lot of hype.
Brett (04:05.298)
and they come at it from a bunch of different angles and a bunch of different teams using a bunch of different tools and there's just no coherent system underneath how they're managing like this growing AI capability, right? And so that that that has some serious challenges and I think you guys are are seem seeming to step in to say, hey, how do we actually bring this into kind of a centralized IP, you know, resource for the organization that kind of compounds and improves over time, right?
Fern Potter (04:18.723)
Yeah.
Fern Potter (04:28.578)
Mm-hmm.
Fern Potter (04:33.196)
Yeah, absolutely. It sounds like they're the kind of people that we're talking to. It's like, where do I start and how do I make sure it's successful off the bat as well? What's those measures of success and how do I not only just make sure it's transforming my business, but also how is AI transforming people and knowledge within the business? So, yeah, it's been a wild few months actually coming off the back of a multi-local and working across the curation space, which is when we first met. So prior to that,
Rio (05:00.773)
Yeah. Well Fern, we knew when we met you at the at the dinner and we like loved hanging with you. We knew you were gonna be super fun and we from remember back then Brett and I were talking, we gotta get Fern on the pot, it's gonna be great when she is so Yeah.
Brett (05:10.298)
Yeah. There was some talk about karaoke. I mean there was I was introducing her to bands from my past. She's like I had never heard of them. English bands. so it was it was a fun it was a fun time.
Fern Potter (05:18.776)
Yeah.
Rio (05:21.231)
House knows a lot about music, there's no doubt. so but we but we were excited to get you on the pod, so we're happy we can finally make it happen. And I guess the the premise or theme of this this discussion we were like decided to dig in with you on was that the AA market's super crowded, right? As you mentioned, there's a lot of hype. companies are promising lots of things, everything from autonomous systems that can run everything in your house to like
Fern Potter (05:23.469)
That's it.
Rio (05:46.309)
Autonomous systems and a and swarms of agents that can even run or transform or change marketing. And we thought the name of your company was cool because it really seems to be based on the premise that of intelligence assistant, right? The machines are powerful, but you need you need auton like autonomous systems that run everything. There need to be people behind it. You chose a name that gently pushes back on that, on the entire premise, right? So it's
You left a senior role, you're at a new firm, you're deeply embedded in there, working with some really like smart co-founders. So we'd love to talk about like what channel what problems are you trying to solve and what does it mean for the future of how how companies and people are using AI? I think it's gonna be a really fun discussion. So starting off the from the beginning, you left multi-local, you were there for a while, and you were chief strategy officer. I've listened to a bunch of your stuff online about curation, and you know, which was a which was a big trend for a number of years. It's funny, we did have someone on the pod recently.
compare curation to I think subprime mortgage backed securities. So so so I maybe some, you know, different opinions about curation in the industry, but but yeah yeah.
Brett (06:45.946)
Mm-hmm.
Fern Potter (06:46.328)
Wow! I need to listen to that one. How interesting.
Fern Potter (06:54.286)
There's a lot of conversations around it.
Brett (06:55.557)
Yeah.
Rio (06:59.782)
But why did you what made you wanna like ex take on this new challenge going from a from a big role like that?
Fern Potter (07:05.334)
Yeah, it's a great question. curation was something I was passionate about whilst I spent over 20 years in media agency, know, the forefront of like social search and then programmatic when it came through. And myself and actually one of my co-founders, James, we worked together at Dentsuit and we built a trusted marketplace. And the trusted marketplace at the time was a mix of partners that you'd work with and broadcasters and media owners that you'd work with directly on the essence of the principles of a brand.
So take, for example, a D'Aggio or a Jaguar Land Rover, a Philip Morris, you'd bridge that into a wonderful, these are the principles of how we operate as a business for every client. And that included all the programmatic buys, all the vendors that we worked with and all the broadcasters and publishers. Trusted Mop, and that was, that was why I loved curation so much. It was about bridging the brand and the media owners back together.
whereby the actual technology from a programmatic perspective had disintermediated a lot of those relationships, but also really taking us onto, instead of cherry picking impressions, getting as much scale as possible. So when I started at MultiLocal, and still I'm super passionate about the power that curation can have, because it then enabled the technology to replicate that model and do it at scale. So you could have a trusted marketplace, but...
Brett (08:02.991)
Yeah.
Fern Potter (08:23.114)
a huge amount. You could find these publishers where audiences weren't necessarily being discovered and you could supersize that trusted marketplace at a brand, product and client level. So it's all about curating those audiences in a meaningful way aligned with the principles of how brands want to operate. Because what we forget
Brett (08:40.431)
Yeah. And w and when you say principles of how brands wanna operate, can you define what that means? I mean, 'cause it sounds like a brand and sort of mission vision sort of strategy. This is who we are, this is how we act and behave and how we communicate with our audience, right? Or audiences. Is that is that is it that foundational, fundamental? Or is it something different?
Fern Potter (08:46.466)
Yeah, absolutely.
Fern Potter (08:50.574)
Yeah.
Fern Potter (09:02.304)
Absolutely. Yeah, I think, and this comes back to links a little bit into what we're doing today, It's strategy, it's planning, it's understanding the brand and the consumer relationship and translating that into the programmatic ecosystem where we turned a bit binary in the way that we were working. We'd forgotten about that. And so when you talk about the principles, it's quality, it's age gating, it's attention, it's ethical practice within data.
Brett (09:17.168)
Yeah.
Brett (09:28.132)
Yeah.
Fern Potter (09:28.814)
There was all of those elements that you can bring together that we kind of lost along the way because we went from knowing where we're buying to knowing who we're buying with.
Rio (09:39.067)
Was it was programmatic kind of what did that? Because one thing one sh one change I noticed was like media shifted to like the people actually spending most of the time buying media, it became very operational people in platforms. And it became this layer of kind of like doers and traffickers and people like within the agencies. And then I I've even worked with a bunch of big media
Fern Potter (09:49.302)
Yeah, exactly.
Rio (09:59.283)
you know, big walled gardens and other media sources. And like their contacts are with those people. I thought that was so crazy. They weren't even talking to the planners as much anymore. The people the people come with the Vision is more like very operational and platform. So it's a that's actually an interesting point.
Brett (10:12.707)
Yeah, and if you and if you have sort of the brand or agency, the buyer, the publisher, and the media provider, where does the consumer fit in that sort of three legged stool, right? 'Cause the consumer's like, these are the cohorts that you're speaking to and that you're reaching in different environments. You better make sure that that's right, otherwise you're gonna you're gonna ruin your chance to develop some sort of relationship, right?
Fern Potter (10:20.34)
Wait second.
Fern Potter (10:32.942)
Yeah, you're absolutely right. this is what actually infuriated me. And I've worked in programmatic for while, so I can say that it's about programmatic. It became so much about the tech. And people forgot about, you know, even a supply chain path.
Brett (10:44.218)
Yeah.
Fern Potter (10:48.35)
people don't put the brand or the consumer in it. Well, the customer is reading the publication that is generating the signals that is then actually enabling the brand to advertise to the customer. It's a circle. It's not a supply chain. And the customer was never in it. And, you know, my experience.
was outside of the trading desk. We built programmatic strategy, programmatic buying units that were fully transparent within the agencies to align with these principles. the relationship was all about how the customer had a relationship with the brand and the or broadcaster, where they resided online at that point in time. And you could get hyper-personalised with that. You can still count the US GDPR over here, not so much, but...
That was the whole reason I loved it. It brought that methodology back of actually thinking about the consumer, thinking about the signals you've got access to and understanding the brand experience that they've got online in a way that, you know, I don't want ad tech to stand up and talk about CTV creativity. It needs to be about people that understand how light TV can influence consumer decisioning and therefore then how you buy it. So I think.
Brett (11:45.785)
Mm.
Fern Potter (11:57.109)
you know, get off my soapbox slightly. think that's what was missing. And that's why, you know, and I still do heavily involved in working with multi local, and that's what curation could achieve. Now, in any ad tech, you're going to get complexity, can get bad actors, you're to get people skimming, you're going to get a lack of transparency.
Rio (12:14.554)
Look at supply path, there's a lot of that, yeah.
Fern Potter (12:16.428)
Well, exactly, exactly. you know, it's the same with everything that when there's these reinventions or evolutions of how we buy and why we buy in that way, sometimes people are great and sometimes they're not. so I think, you know, again, there's a lot of if you don't work with the right people that won't lift the hood and go, this is how we're doing it. And this is how we're doing it for you. Then.
don't work with them because they're not being transparent. there's a lot of really positive conversations, results and performance that comes out of working with curators. And I think the way that multi-local work was really different is curation as a service into business. You see all the agencies building their own curation executions and teams now. We could just plug and play into business. And I think that was, it's a real strong proposition.
And I just Stagwell, think last week announced that they were building there. So there's as much as it was not as noisy as it was because I think AI and supply side decisioning and containerization has slightly taken over. I'm still obsessed by all of it, by the way. There's there's there's curation still happening. It's just the news cycle shifted, I think.
Brett (13:09.508)
So
Brett (13:24.334)
Yeah, we can tell.
Brett (13:31.822)
Yeah, it's not a subprime mortgage, is what you're saying.
Fern Potter (13:34.286)
100 % no.
Rio (13:34.888)
Well well it doesn't have to be. I I think it depends on who's curating, right? I mean, if you have crappy inventory that you're mixing together, it doesn't make it any less crappy, right? Even if you attach a DL ID to it and try to you know, 'cause th that was the thing about mortgage backed securities, right? I mean, they were taking all of these mortgages, they're bundling them together, and all it took was one or two bad mortgages in the in the in the the bundle, right? To cause the entire the curated bun to cause it to be worthless. So so so I mean that's
Brett (13:42.018)
Yeah. Yeah.
Fern Potter (13:50.158)
Yes.
Brett (13:53.615)
Yep. To cause the whole to to threaten the entire thing. Yeah. It's it's sort of the dang it's the danger of averages, right? You know, and if you're if you're bundling in a bunch of MFA and terrible media content into kind of what you're calling a premium curated category, that's it's it's yeah, you're kinda hiding the truth in a way, right? 'Cause only a certain percentage of that of that that group of publishers might actually be premium.
Fern Potter (14:03.138)
Yeah, yeah.
Fern Potter (14:16.418)
Yeah, exactly.
Fern Potter (14:20.706)
Yeah.
So you need that transparency, but the same is kind of happening with AI, right, to your point. It's like AI is like the sprinkling the glitter, rolling it around and making it look shiny. I mean, if what you put in isn't good, if what you are building isn't of value, then no amount of putting AI on top of it is going to make it any more useful to the people that are buying it. So I think there's watch outs there, but yeah, curation is definitely still a topic and it's evolving, which is great to see.
Brett (14:49.966)
Yeah. So so curi cu
Rio (14:51.335)
Looking at looking at AI though, like really quickly then, like umbrella, did you if you had another question on curation, let me why don't you ask that, then I can go to
Brett (14:58.54)
No, no, it was more about I I was more about 'cause we've jumped right into the heavy topics and I and I thought, hey, I'd like to hear a little bit more about your founder story and kind of what brought you what made you 'cause I've taken the leap myself six months ago and it's been the zero to one challenge. signal and noise, Rio and I run on the side. and it's it's
Fern Potter (15:08.855)
Yeah.
Fern Potter (15:18.253)
Yep.
Brett (15:20.322)
It's a fear of death and sort of you have to be willing to deal with uncertainty, a huge amount of uncertainty and a lack of structure, and you have to you have to be a builder and and it takes industriousness and sort of fearlessness. What what what made you take the leap and become a founder and found this company?
Fern Potter (15:26.637)
Yeah.
Fern Potter (15:40.939)
Yeah, it's such a good question. Thank you for sharing those kind of thoughts and feelings as well. I've definitely cycled through, you know, daily. I just wrote it. It's like you go from this like hype to existential crises on like an hourly basis. you know, me and what
Brett (15:44.656)
I I yeah, I I I just I cry on my bed all you know, in the fetal position on a nightly basis.
Brett (16:01.263)
Yeah.
Fern Potter (16:04.43)
I mean, love technology and data always have. So we've had hands in hands in what AI is doing and how it's working for several years. And I think it was serendipity with two co-founders. And I think that's something, know, just mentioning those emotions, you've having the right people at the right time to go, yeah, let's fucking do this. And we're all so resilient and we all have such great background and relationships together. It was just what God is here won't get us there.
Brett (16:22.244)
Yeah.
Fern Potter (16:32.662)
in any anything, in any state of our life in business, know, AI is fundamentally changing not just business, but everybody's life. It's impacting everything that we do all the time, you know, so it's mass consumer change. And I think that intrigue and curiosity as well as desire and having the right people around you to go, let's go, was what made me know that. Absolutely. Yeah.
Brett (16:43.055)
Yeah.
Brett (16:53.476)
Yeah, you you didn't want to stick around in a company and not and miss the wave. Was that part of the the Yeah, 'cause I felt I felt that. I'm like, I could stay here for another three or four years at the in the twil twilight of my career, if you want to call it that. and then just exit four or five years from now, no kind of better off than where I started, maybe a little bit better off, but I f but I felt there was this this definitely this fear that that I'm gonna miss out and I wanna be involved in this, right? This entire AI search, n being a user, a builder
Fern Potter (17:05.134)
Ha
Fern Potter (17:19.426)
Yeah.
Brett (17:23.064)
It was sort of a it's kind of a personal and technical transformation. It's like a you know, almost changing like I move towards like I I've gotta stop thinking of myself as a specific function, right? And limiting myself to that function and start being much broader and deeper across a a bunch of different which is kind of what you have to think about when you're becoming a CEO or a founder, right? You're you've got multiple areas that you need to to manage, right? So is that has that been part of your journey as well? I mean, is that does that resonate?
Fern Potter (17:32.461)
Yeah.
Fern Potter (17:42.68)
Why? You've got to do all of it. Yeah.
Fern Potter (17:49.583)
Yeah, think I can definitely relate to that as well. There was this fear, but also this immense opportunity, and perhaps that we've not seen before. And I think between the three of us, so we've all worked together in some guys previously, there's over 70 years of strategy and planning between us, not that we'd like to admit to that too much. there's James who, he's
Brett (17:56.047)
Yeah.
Brett (18:15.096)
Rio and I just together have sixty, seventy years of experience. So we we're older than you.
Fern Potter (18:20.333)
You
That's fair, that's James, an engineer by trade, he coded, you know, that's what he said. He's founded different media businesses as well, and then became planning and strategy. know, Sarah's worked in planning and strategy with brands, but also is a phenomenal force in building B2B SaaS businesses. So you've got great credibility there. And then myself kind of bridging that technology data and commercial gap as well. So I think the three of us also were like, okay, actually this...
We've got some natural complementary skills that fit together. Let's give it a red hot go.
Brett (18:53.496)
And then she gotta figure out who's gonna be the CEO.
Rio (18:55.504)
Yeah. Well and by the way, Brett's exaggerating a little bit. So we met in two thousand, which is really our first media job, right? So that so that's like what? Twenty twenty six times two, right? Fift fift fifty two years, right? It suddenly suddenly became sixty or seventy, right? That's like this is I this is not adding up, right?
Fern Potter (18:57.824)
I don't wanna be an adult.
Brett (19:03.034)
We have fifty years of combined experience. Yeah. We're not quite that old, right? Yeah. I d I told I t I told one of our previous guests that I had a nineteen year old and he about fell off his chair. He's like, What did you have kids when you were when you were like in your early teens? I'm like, I don't think you know how old I am.
Fern Potter (19:04.558)
Rio's trying to crawl it back now. Rio's like, don't want to be a placeholder.
Fern Potter (19:18.254)
What's this?
Fern Potter (19:24.526)
I think that's a compliment, that's wonderful.
Rio (19:28.402)
But Fern, looking at artificial intelligence, like your company argues, looking at the some of the materials you've put out, that artificial intelligence is not actually intelligent, and I'm using quotes here in the in the human sense, and that it it's incredible at recognizing patterns, doing complex tasks, doing them very quickly and very efficiently, much more than people can, but it lacks context, consequence, and lived experience. Why is it do you agree with this and why is this
distinction commercially important and how does it relate to what your company's doing?
Fern Potter (20:02.168)
Yeah, great question. We've got a real simple process, think, build, adopt. And the think part is really important because that's when you set up what a good output looks like. And at the moment, with all the AI tooling that's out there, not just the frontier LLMs, you don't know what you don't know. And so unless you have a really clear...
understanding of what you want to get to and why you want to get there. You've got no strategic alignment, no commercial alignment and no stakeholder alignment in your business. AR can't do that for you. So bringing that, that's where that human intelligence is so important. You've got to bring that knowledge base, not just from a technical perspective, I'm not trying to plug it in data. Yeah, that's really important. But you've got to bring that knowledge base of your business, the understanding of your customer.
Brett (20:34.125)
Yeah.
Fern Potter (20:48.91)
into what you want to get to and where and why and what success looks like when you get there. But then also understand iteration. Also understand that you're going to get to this point and then you're going to grow from it. And again, running a whole business around that is really important. And so that's where it's not smart. You can't just and we've seen this time and time again go, right, I need AI, let's go. Or a CEO has come in and said, right, everyone in the business needs to use AI with no clear objective as to what, where, why.
Brett (21:02.211)
Yeah.
Fern Potter (21:18.864)
and how you're going to get there, or even that psychological safety that we've talked about of aligning businesses around how my job is going to change as I do it. we do argue that. We fundamentally believe our philosophy is all around human first and human last, not just at the think stage, but also the output. You see a lot of businesses.
building things that sit over here in isolation, another platform to log into, another dashboard here, another KPI, something else to use. Our strategy and planning background is you meet a customer where they're at.
So even if that's working with internal teams or working with their customers, it's how do we make this experience as frictionless as possible for usage? So you train on adoption, but you make the, what if we call them engines, built to last and sustain business operations. I think that, yeah, we're fully in on that philosophy because it is about assisting the intelligence of humans, not replacing them.
Brett (22:18.957)
So so what
Rio (22:19.216)
Well it's I think it's been a dirty little secret in consulting anyway, that a lot of the AI work has really just been training people how to use AI, right? I think Accenture, these big numbers a couple of years ago, about billions of dollars. Then when you dug in it was actually big contracts with government agencies, like quite literally telling them how to like use Chat GPT and like open a prompt. which was you know, but I'm not saying that's not needed, but y you know, he's not really
Fern Potter (22:40.822)
It's true. Yeah.
Rio (22:41.636)
Is that really AI work? Was kind of the question then? So and but at the same time, a lot of what we're seeing, like to enable AI, you do need fundamental change to the way organizations are structured and operate, right? So how much of this of AI work is really change management? How much of it is old change management that's not very valuable versus like really fundamental structural and operational and functional change to the way the companies do business?
Fern Potter (23:10.102)
Yeah, it's a great question because everybody is at a different stage and level of maturation as to where they're at. You know, we as a business build in lightweight product. As I said, we call them engines. It's about assessing where you are, where you want to get to, what you've got in place now and what you need to be able to complement that change. But the way in which we operate is to build a lightweight package as such, as I said, we're SaaS business that can demonstrate value.
both to the business and to the outcome, whatever we've aligned on that being, and then build from there so that businesses and founders feel confident in investing their capex in the right place. And that hasn't been done to date because when you come in and say, right, actually, you need to change the whole infrastructure and all your data architecture that exists beneath your business, but we're not going to be able to tell you what fundamental shift that's going to have to your bottom line.
then that's where the uncertainty and the unconfidence and the lack of adoption comes in. Also, there's just a shit ton of complexity out there. It is complex, but it shouldn't be too custom. The way we talk about it can be really complex. I think you need to work with partners that cut through all that to go, honestly, this is where you're at and this is where you want to get to and this is how we do it.
Brett (24:31.051)
Now to g yeah, now let's get let's get specific about like sort of what you're dy 'cause it 'cause it you this could have so many different applications. And one thing that I'm seeing is a common trend within within organizations is is and it's also becoming kind of a a risk to brands adopting AI. and we're having a conversation in one of our upcoming episodes with Eddie Drake from Snowflake to talk about some of this, is is sort of the IP threat, right? And we won't get in and the idea and and I'll and I'll tie this back to what what I think
Fern Potter (24:37.838)
Thanks.
Brett (25:00.109)
you're getting at from a from a build perspective is the idea that brands fear you know, they have a ton of IP that's spread across a ton of different platforms, a ton of different people, a ton of different communication mechanisms, email, Slack, Zoom, granola, whatever it might be, and very off more often than not, that information is not curated, gathered, and decisioned upon.
Right. And so you so you just so it so it actually hampers the the sort of next best action ability of the organization because they're usually making decisions either in silos or it requires a lot of like sort of matrix, you know, large meeting environments where people are trying to kind of bridge the gaps between organizations and in and and data access. Not everybody has the same level of data access or the same information availability, right? So it seems to me.
Long story short, that a lot of the the push to become AI native as an organization is kind of falls into this sort of the I hate to use a a a hackneyed term, but this digital twin notion. The idea that you can actually build a an ecosystem that lives or or or a knowledge graph, let's say.
That lives within your environment as a as a as a brand and it's headless. So it actually you can actually send out agents to basically pull in all the relevant information across all of your disparate platforms and teams to create some sort of knowledge base that you then can use to kind of democratize data access to help better decisioning. Right? Is that kind of structurally how you guys are thinking about it for helping organizations? Maybe that's too complex, but helping them become AI first?
Fern Potter (26:41.92)
Yeah.
Brett (26:42.016)
And leverage sort of like that compounding intelligence which has kind of you know, is that how you guys are thinking about it from an architecture perspective?
Fern Potter (26:48.046)
Yeah, think you can use it as like, as I said, we use this framing of intelligence engine, which is agents, it's automations, it's API's, it's your server instance, it's your knowledge base. It's being able to process all of that information into really simple to use valuable user interfaces. So to your point, let's go into some examples to help kind of picture that. We built a strategy and planning tool.
Brett (26:58.572)
Yeah.
Brett (27:10.946)
Yeah, yeah. Illustrate this. Yeah.
Fern Potter (27:17.73)
that can actually consolidate the time it takes to brief respond by to your point creating and reducing, creating a knowledge base, but reducing all the information silos that exist across the business. But then being able to democratize ideation because everyone can go into the platform as opposed to, you know, there's five people in a room that we could afford to fly out to some headquarter.
Brett (27:18.051)
Yeah.
Brett (27:31.309)
Yep.
Brett (27:38.572)
Yeah. So they have equal they have the access to the same level of information, right? That's how you yeah.
Fern Potter (27:42.145)
Absolutely. So you've got complete equity, you've got speed. So what happens there is you've got the opportunity to answer more briefs with higher quality, higher consistency, more rapid rates and actually spend more time on focusing on customer relationships and being able to, you know, into it.
Brett (27:53.485)
Yeah.
Fern Potter (28:01.346)
and use human judgment on what the outputs are. So those systems exist, but on the flip side of that, we're also working with small, medium businesses. So like that SME space that actually haven't got any of this, that are not really thinking about...
the knowledge element of it, but really thinking about their own customer. And like we've built a customer service agent that is replicative of the brand and the information within the brand for an insurance company that has enabled them to frontline their customer service without it going through to their sales team and improve their lead conversion rate by 25 % because their sales team can handle, you know, all of those sales conversations as they should. And the customer service agents are handled
handling all of the questions and frequently asked questions. And that's a voice agent. It's a person that sat there. So there is multiple applications of how this can work, but everything is coded inside somebody's system. Everything should have lived in their ecosystem, the endpoints of their ecosystem. So I think, you know, and we're really
Brett (29:04.844)
Yeah. Yeah, it it lives within their own environment, right?
Fern Potter (29:13.952)
a lot of conversations around the barriers and the intellectual property. I think you've got to have that flexibility. the responsibility needs to come from having human judgement at every stage to understand exactly where information goes, what it does and what the output of that is. So we do force a lot of gating within our products to make sure that everything is captured.
Rio (29:39.444)
So for like I one point you made a second ago, I really liked it was the combination of the power of AI, the power of large language models, but then the domain and industry and specific and maybe role and company experience that could be in their internal systems, it could be in documentation, or it could just be in their heads, right? I think that's really the power of this. That yeah, you can get a lot from the large language models that the hyperscalers are deploying. They're incredible, right? And even some of the vertical LLMs that are being published. But I think that
When it when where the rubber meets the road, where you're deploying these at a company level or an individual level, it's a combination of that experience, getting it out of people and out of companies and out of teams, and using it to build build this application layer on top of the the foundational LMs. That's where the power really emerges. I'd love to see if you agree with that and like maybe give us some examples of how you've done that.
Fern Potter (30:32.544)
Yeah, I do often talk about the fact that actually people are the product when it comes to AI and like AI is just the accelerant of the fuel through a system. Because it's not as it to your point, it's not just about the human intelligence that goes in and the thinking but all the signals and all that information and you know, the experience that you've collected over time, if you can codify that into a system that then is democratised across the business, it's like fucking awesome. That's you know,
Brett (31:00.704)
Yeah. The the that is the promised land and that's what a lot of the AI companies have talked about. It's exactly that. It's it's sort of the people intelligence and temporal intelligence and how do you actually codify that and capture that in yeah.
Fern Potter (31:01.548)
That's your brand, your... it's beautiful! Yeah? And it's absolutely...
Yeah, it's true. It stops AI making everything commoditized as well. Sea of saneness. You've got no It's
Brett (31:17.185)
Yep. Yeah.
Rio (31:18.786)
And the same.
Brett (31:21.11)
Yeah. Well and and you're also but you're also developing and I'm then I'm gonna make an assumption here, but you're put you're developing a system that I'm assuming is agnostic to model usage. Meaning meaning it's compl you know, you can go to, you know, it c it can leverage a a large language model that's, you know, kind of a frontier model, it can leverage other i emerging models, and it can it can prioritize which model to use based on the the the complexity of the tasks so that you're so there's a cost and efficiency play here.
Fern Potter (31:31.694)
minutes.
Fern Potter (31:42.071)
Yeah, but.
Fern Potter (31:45.761)
it does.
I'm so.
Brett (31:48.833)
Right, in terms of which model you use when based on which tax task and that's gotta be automated. And it also re you know, back to that IP point, is it reduces the fear and the concerns that brands feel that hey, we're putting our information, our our people's intelligence and data intelligence, whatever, into these other places and and giving away our IP, right? And they feel like they're gonna be disintermediated at some time some point down in the future, right? Is that a big concern that you hear from clients going like where is this going?
Fern Potter (31:52.375)
insight.
Fern Potter (32:06.636)
Yeah. Exactly. Exactly.
Fern Potter (32:18.509)
Yeah.
Brett (32:18.52)
How does this how do we b have governance around this?
Fern Potter (32:21.454)
And to your point, need to, and what's happened today is exactly that. Okay, well, you know, when the directive has been going, use AI and see what you can do with it. Like no one's actually thought, well, have we got paid subscriptions? Where's that data going? Has people, you know, people turn the right settings on and off to make sure that this isn't bled out. And so I think.
We've moved on from that into how do we create internal systems using the LLMs. You've got these amazing amount of suites of tools and specialisms that are built on top of them. And then we're kind of here going, right, us build that into an engine for you, choosing the right models at the right time in the right tech. And the model thing is really important. mean, there's so much fragility in AI. You don't know what's going to happen in two months time. know, Fable came out and then went again. If you built a whole business on a singular LLM,
Brett (33:04.269)
Yeah.
Fern Potter (33:12.142)
That's going to be tricky and tricky to scale, especially in multiple markets.
Rio (33:14.991)
Risky, yeah. And it might then they they might also disintermediate you by launching launching a product that's exactly like yours, right? There's there's a lot of stories about that kind
Fern Potter (33:20.798)
absolutely. Yeah, yeah, that's true. mean, it's brilliant how they are evolving so quickly and this speed is unprecedented. think think AI is a cultural shift as well as a technology shift.
I liken it to social media in that way because it's fundamentally changing how we receive information, have conversations, et cetera, et cetera. But the speed is unprecedented. We haven't seen anything like it. You've got to have so much resilience as a business built in, as well as so much resilience as an individual within a business to be able to...
Understand the risk, understand the compliance, understand the amount of opportunity there is. Work with people to cut through the complexity, though.
Brett (34:12.514)
Yeah.
Rio (34:16.083)
So for an you said the AI lacks context nuance and an a true understanding of actual commercial risk, which I don't disagree with, by the way. Assuming these things are true, like do you feel companies are moving too quickly, like y having AI actually run things or deploying autonomous AI agents? And what would be your advice to any companies considering this?
Fern Potter (34:39.104)
my gosh. So I had like the this fangirl moment apart from obviously our conversation here. I did have the pleasure when I was at Multilocal of talking to Shelley Plartmer who of Pong, he's incredible voice on AI and he's ace and we we got talking about autonomous, you know how
and where we're going to get to. And I think I have no doubt in my mind that that is going to happen. Like in everything that we do, there's going to be so many autonomous systems. But I kind of liken it to, you know, the autonomous cars like catching a Waymo.
I caught Waymo in Austin when I was there last year. was like, this is novel. This is wild. Like, how is this doing it? You know, and taking photos and videos, show my pals, like, I'm in a driverless car. Do I want that every day? Absolutely not. Would I trust that on any kind of UK road system? Again, absolutely not. We've not got the grid system, Phil. So I feel like you're going to get there. Are we there yet? Perhaps some people are. We are.
Brett (35:23.373)
Yeah.
Rio (35:41.275)
That's the full self driving's kind of amazing. I mean it's gotten so good recently. I actually use it most of the time. on it it can go eighty miles an hour. It's it's incr it's it's it's really incredible how good it's gotten the last few years. But sorry, I cut you off Fern.
Fern Potter (35:44.462)
Brett (35:47.168)
Yeah. I've I've been on a Tesla. That is. Yeah, it is incredible.
Fern Potter (35:47.212)
You're it, Lou.
Fern Potter (35:54.677)
No, you're fine. I'm just, I'm such a passionate driver. I love the actual, actually driving cars that I just don't think I'll ever be able to do it, but maybe that's not going to be my choice.
Brett (36:03.405)
Do you drive like a do you do 'cause you live in the Shire or do you drive like a little car on little streets? Right?
Rio (36:08.925)
The Shire. They they have cars in the Shire?
Fern Potter (36:12.278)
No, we just, we just sort out the wheel.
Brett (36:14.157)
Yeah. Nobody wears shoes. Right?
Rio (36:15.933)
Ha ha ha.
The ride horses, Brett.
Fern Potter (36:20.558)
Brett, I've got this picture of where I live in your mind needs to be altered immediately. We're doing a group. Ria's going...
Rio (36:24.355)
Yeah.
Brett (36:25.469)
I am just thinking it's it is it is the Shire from Lord of the Rings.
Rio (36:28.328)
They live in they live in they live in caves, Brett, in little hobbit holes.
Brett (36:31.757)
No, but there are little windy streets. Little windy hilly streets. Yeah.
Fern Potter (36:33.528)
Thanks. There is more industries, but no, I have a very powerful vehicle.
Brett (36:41.023)
Yeah. He's like I do have a BMW.
Rio (36:41.031)
No, but try driving for driving, I love driving. Driving is so fun. I love I mean, I had a B and W for many years. Driving's fun. I I learned how to drive on a stick shift. I love driving. I love driving fast. But like full self driving is incredible.
Fern Potter (36:45.612)
Yeah, enjoy yourself.
Fern Potter (36:50.478)
It's been a... up in the trash.
okay. Well, yeah, this is one we'll have to agree to disagree on. But I can imagine. So back to that point of some people are ready for autonomy and full autonomous systems, others are not. And I think it's definitely it's coming, not just the cars, but it's coming in decision making, it's, you know, in the way that we operate. And I think I definitely applaud it in some areas. But, you know, for us, the businesses that we're working with are in the start of this journey into how do they use AI to
Brett (37:07.467)
Well
Fern Potter (37:23.958)
to speed up and to increase productivity and drive growth. And that's got human gating at every part. In fact, we built a tool the other day and someone said, can you take the human gating out? And we're basically, we're all kind of sat there going, should we, can we? Does that actually go against the integrity of the way that we're working? But clients are always right.
Brett (37:46.54)
Yeah, the well and I mean McKinsey's done put out some re there's been a ton of research on this that that that I think a third of all ki only only a third of companies, sorry, sixty six percent of companies are not scaling. They're really in they're kind of stuck in pilot mode and experiment experimentation mode.
Fern Potter (38:00.504)
Yeah.
Rio (38:02.611)
I don't know, a lot of those studies that Gartner had the one where ninety five percent of pilots fail. I think that that I mean that may be true. I but I think a lot of these statistics are just to try to scare.
Fern Potter (38:07.246)
I don't know yet.
Brett (38:07.869)
No, but it but point but
Well no, but you're talking about autonomy and and and Fern is talking about the most basic use cases of sort of saying we're going to set up just y and you need to start at the simplest level in order to ensure that you have data security, governance, people buy in, right? Start with the use case and build from there. It's it's a much simpler starting point than I think you know, the this notion of like we're you know, we're gonna have autonomous agents running organizations.
Fern Potter (38:26.926)
Yeah.
Rio (38:30.803)
Yeah, it's all important, yeah.
Fern Potter (38:39.438)
I mean, there's definitely businesses that are there, but so many of them are not. There's a huge market in the SMEs of how they can use AI to help grow their business. And I think you need to be able to democratise those skill sets across all businesses. And I think it's funny that...
Brett (38:59.913)
And SMEs, you're not talking subject matter expertise, you're talking small and medium enterprises. yes, because we typically call them SMBs. but I just wanted to make sure. Yeah, I was yeah, I was yeah, I figured you meant that, but the small and medium organizations that are trying to leverage it for for kind of Yeah, for yeah, and they say Nike they say Nike instead of Nike and Hyperbole instead of hyperbole.
Fern Potter (39:04.952)
Mm.
Fern Potter (39:09.16)
SMB, I know, I use SMB and SME interchangeably. Yeah.
Rio (39:16.187)
So they called them in a shire Brett.
Fern Potter (39:19.726)
I never could be with this doubt.
Rio (39:25.425)
And a Adidas instead of Adidas, right, yeah. But that is actually the right way to say it, but sh
Fern Potter (39:25.526)
I feel like you've just taken the piss off of us now, guys.
Brett (39:29.671)
Aubergine instead of eggplant.
Fern Potter (39:35.031)
Fuck!
Rio (39:35.429)
Aluminium. That's my favorite.
Brett (39:36.333)
Sorry, sorry. We just finished an hour.
Fern Potter (39:38.646)
I
I think you need, next time you're over, we're doing a road trip, not in an autonomous car. And I can mystify this picture that you've put together for all your listeners of the Shires. But yeah, or maybe not so much.
Brett (39:49.514)
Yeah.
Brett (39:55.465)
Yeah.
Rio (39:56.594)
No, but we but I we we both are huge fans of London and of the UK generally. I mean I I love Manchester too, great town. I s so I I definitely enjoy my trips there. the food's gotten better the last twenty five years.
Fern Potter (40:00.911)
Good. Good. Yeah, yeah, it's not bad. It's not bad.
Brett (40:09.973)
Well, you why they you why they say and we'll we'll get back on topic, but they say the food's gotten better because the English aren't cooking it.
Fern Potter (40:15.982)
As someone who's probably cooked a meal as sophisticated as toast her whole life, I'm going comment on that. I that's about time to cook.
Brett (40:21.946)
Yeah. A little marmite on that toast?
Rio (40:24.921)
English like traditional English food food like roasts and vegetables are pretty good actually. But I wouldn't I don't know if I wanna eat it all the time. But but yeah, the curries are incredible. I love a fish I'm a huge fish and chips fan. Much better over there.
Brett (40:28.876)
Yeah.
Fern Potter (40:34.604)
Yeah, that's good. Yeah, we do. There's some pockets, pockets of excellent, think definitely improved. I agree with you, Where I'm lost now, I mean, maybe cooking is the autonomy that I require for sure.
Rio (40:43.709)
But get
AI cooks, right? Maybe is the thing there. But but looking at like AI being deployed within companies, I mean human in the loop has been a a bit of advice that most people I work in I that work in AI have always said is very important before we get to full autonomy, right? Making sure there are the right checks and balances. So when you're building these models, like
Brett (40:49.323)
Yeah.
Rio (41:12.379)
Human first AI, meaning there's human does that mean there's humans in in the process or does that mean it's human friendly? Like what do you mean by that exactly? I know that was something that you've you've called out quite on your website materials. Human first AI.
Fern Potter (41:25.654)
Yeah, again, it comes back to that kind of, you know, thinking power and people element of what we do. So it's a simple process in the way that we work so that, you know, would keep things straightforward. One definitely one of the things I've learned from Antec is that the complexity just creates so much ambiguity that it becomes unnecessary friction. So it's throughout everything that we do. So the think phase again, is actually getting all the key stakeholders in a room to align on what
Brett (41:45.708)
Yeah.
Fern Potter (41:52.205)
what good looks like, what the output needs to look like and who's doing what within that project. We go and build and that's gated. That's done as sprints are. The sprints are a lot quicker as we move faster. And the adopt phase is all about people. It's about the training of the individuals, but it's also about thinking, how do I deploy this in the most...
frictionless way possible. So what I mean by that, and I think we talked about it before, I don't want to log into another dashboard. I don't want to have to go to these URLs. These businesses have systems that exist already. We're not replicating, we're not building additional operational workflow. You can deploy within the systems that exist. And like that comes back, Brett, to your point about the intellectual property piece as well. It has to be. I think the...
Brett (42:32.705)
So it's h headless. Yeah. And and do you give do you give the option do you give the option of interacting with with the capability, the let's say it's particular project, right? Like w either through their own systems, like it could be through their own claw.
application within their organization or through a UI where they log in and and you know, whether it's conversational or otherwise, I mean are you giving them the option to go either way? Like you could you could actually speak to this through Slack, through Claude, or you could come through a a UI. Is that is it that sort of flexibility?
Fern Potter (42:55.533)
What?
Fern Potter (42:59.276)
Yeah, a huge amount.
Fern Potter (43:05.576)
Exactly, so huge amount of flexibility. We deploy in Claude, we've deployed in Chrome extensions, we can deploy... I mean, for a university, we've deployed in Telegram because that's where the students were hyper-indexing of using Telegram as an app. So, there's various different ways, but that's just, again, the joy of the interoperability you get with the technology now, that you can ensure the technology doesn't get in the way of itself.
Brett (43:18.325)
Yeah. Yeah.
Fern Potter (43:35.023)
And that the adoption is again from the psychological, from a change behavior perspective with the people within the team, it makes life as easy as possible. I don't want to just replicate the frictions that exist by sprinkling some AI on it. You want to solve for it. And that's where you're going to get and, you know, get back to the usage and therefore see the success that goes, right, that has hit my KPI. Let's go and keep iterating. And what can I do next? And we get that a lot. We start with small projects with clients.
Brett (43:35.148)
Yeah.
Fern Potter (44:04.968)
And we get the next ones. get the next ones. They're like, yeah, take the capex. Yeah, this has hit my KPIs because you're not doing, and I we talked about this at the top of the conversation. you need to do this massive structural change and go in there and blow everything up because AI exists. Well, do you?
Brett (44:19.275)
Yeah.
Yeah, it's hard it's hard to people yeah, it's hard for people to wrap their head their head around the complexity of a business.
Fern Potter (44:25.742)
It's been the same with any tech. Well, exactly.
Brett (44:27.252)
And the millions of hours of human hours it's taken to build out this multifaceted organization across multiple functions. So so as a startup, you know, and I've I'm certainly in that mode myself, how do you actually go in there and say, hey, this is where we specialize from a use case? You've talked about a couple of use cases. You had a customer success use case, right? How do you actually go in and say, these are the the the core problems that we you know, it's not we're not boiling the ocean? Because I've I've had a lot of conversations with AI companies that kinda, at the end of the day, you're like, okay, I get it, you're creating
Fern Potter (44:37.132)
Yeah.
Brett (44:57.196)
knowledge graph. You're creating this and this and this is how it's architected. You know, but then when I you know and they're telling me the what, right? Over and over and over again. And I'm like, so what about the why? Like why does this exist? And they and oftentimes they have problems articulating it. And they and and it and it becomes kind of a
We can do anything because we've built this agentic swarm system that's that's headless, that can fit that can plug into any stack and dot dot dot read any data. You know, but I'm like I'm like that doesn't seem like a winning go-to-market strategy to sort of say we can if you do everything, you do nothing at all. So how have you guys, is that something that you guys are like sort of racking your brains over? Like, where's our specialty? Where's our focus area? You know, if we want to go raise capital, it's get we've got to be the master of like one domain.
Not twenty-five domains. How have you guys thought about that?
Fern Potter (45:44.12)
Thank you.
Yeah, it's a great question because, we're early days and there is a network effect of going, okay, what the realms of the possible. So there's probably two kind of clear routes to market that we've got at the moment is our background in strategy and planning. It makes sense that we work with brands directly and some agencies on being able to, you know, codify what they're doing from an AI perspective. And then as I said, that SMB space,
Brett (46:10.549)
Yeah.
Fern Potter (46:17.058)
whereby we are building commercial product to help accelerate their conversions within the business, if that makes sense. So there's kind of two streams, but it is really broad. I do agree. You want to be able to go and we're not the jack of all trades, but we're definitely understand where we can laser focus. So I think that will be refined over time as well.
Rio (46:43.004)
So Fern, it must be interesting for you with someone who has such extensive experience in ad tech, in advertising, and in media.
Rio (46:53.57)
As you mentioned, taking a more generalist approach, which I think you kind of have to have at this point, bringing an AR company to market. Some of the use cases you're mentioning, I mean, you mentioned customer service, which is kind of related, but it's not media, right? As well as some of the other things from enterprise planning and strategy. I know you have a strategy background, you were chief strategy officer, but I think I think it must be a trip for you to actually not be focusing specifically on digital media and ad tech and be looking more broadly at solving problems for companies.
Fern Potter (47:22.752)
Yeah, it's been an interesting journey, think, for my for myself from a change perspective. fortunately, James and Sarah have had similar experiences where we're just suddenly like, OK, well, you know, your eyes still slightly on that. But the learnings have been incredible of, you know, the last of the first few years at Multilocal, I was in product, you know, building in ad tech is still phenomenally fast. But
actually going back to the roots. I think this is true of everything that you do. You just got to focus on the customer. You've got to focus on customer because if you build from, you know, and I don't just mean the client, I mean, the clients that we work with, and their customer, what is it that's going to shift the dial with them? What is it that you then need to do? And then why is it you're doing it and bridge that together? And I think it's, I think that's true of everything that
I've done throughout my career, whether it's building brand strategies, I've worked at creative agencies as well. I've built platforms and technology within businesses that have creative experience. And a lot of what we've done within the ad tech space over the last several years has always been to do with data and signals and orchestrations of audiences for consumers to improve the brand experience. that's kind of the thread that runs.
The house.
Brett (48:45.929)
Yeah. And if the p and if the pa and if the pain point happens, if the pain point's a customer success pain point that's causing churn within the organization, if it's a customer acquisition pain point, if it's what you know, communications or brand strategy, you're you're gonna like that's that's the connective tissue here. And you're gonna you can solve a lot of these problems without having l let's say a mastery of customer success 'cause you didn't come from a customer success background, right? Like you that that isn't necessary
Rio (48:46.78)
But but
Fern Potter (49:01.996)
We find them.
Rio (49:10.983)
Yeah. Well but but I do think for like before I mean interject before you jump in, like I do think a good strategist will be able to just solve problems, right? You'll be able to take your the same and and I that's really what this should be about. But I do think like to to have the time to add tech here. Ad tech is an industry that's famous for I think doing two things. Number one, creating hammers and running around looking for nails, and number two would be creating creating insane complexity.
Fern Potter (49:11.95)
Yeah.
Brett (49:19.305)
Yeah.
Fern Potter (49:19.438)
It's just a small problem. Yeah.
Fern Potter (49:26.446)
Mm-hmm.
Brett (49:32.299)
Building it, building it hoping they will come, right?
Fern Potter (49:38.422)
are so many.
Rio (49:38.427)
It's unnecessary, then creating tools to solve the complexity and then selling those, right? And mon and then toll taking based on that. So
Brett (49:41.907)
Yeah. Yeah.
Fern Potter (49:43.587)
Oh, I know. Oh, yeah. He's like, he just said something resonated me. I felt it. The thing about ad tech and a lot of guys and a lot of people do this is like optimize the problem as opposed to solving it and then make money off the back of it. And he's like that that that was a push for me as well, you know, when coming out of pure ad tech. Yeah, because you just.
Brett (49:56.937)
Yeah.
Brett (50:03.573)
To push aga to push against that. Right? I think I think there's an analogy to like our food. You could like think about food. Food in the UK and food in the US, right? You get this sort of like we're gonna pr we're gonna mass produce all of this food that makes people sick, and then we're gonna sell them medicine to help them. You know, I'm not saying that those that those two parties were necessarily in cahoots to say, let's make people sick so we can sell more medicine. But, you know, that you you created a problem that was probably avoidable. Right?
Rio (50:17.427)
Yeah.
Fern Potter (50:22.606)
Yeah.
Yeah.
Fern Potter (50:32.226)
Yeah, mean, just, you know, it's like the fraud stuff, isn't it? Well, just stop doing it. Stop buying that thing. But then you've got technology that Dr. would speak to it better than me.
Brett (50:37.268)
Yeah.
Brett (50:41.161)
Those tortilla chips are so tasty though and those salt and d those crisps Those crisps are irresistible
Rio (50:47.139)
MFAs make your make your metrics look so much better, right? Right, Brett? Right? The junk food the junk food boosts your KPIs, make makes your campaigns look that you're performing better. It's still bad
Fern Potter (50:50.382)
my god!
Fern Potter (50:57.652)
Absolutely, everyone enjoys them at the time and they don't realise the consequence. I forgot the question, I'm just thinking about dinner.
Brett (51:01.417)
Ha ha.
Rio (51:07.273)
but the anal I guess the analogy would be like so that like that I was making a joke about ad tech, but like how do we prevent that from happening in AI?
Brett (51:13.877)
Yeah.
Fern Potter (51:14.584)
God, mean, like, there's a bigger question around that. And it's funny enough, myself and Sarah, my co-founder, we were at the House of Commons last week in parliament lobbying for ethical guardrails to protect women and girls online in AI, because there is none, really. These LLMs are roaming free and the information that you can have access to is...
Brett (51:37.491)
Yeah.
Fern Potter (51:41.312)
is well not only ludicrously biased but also really quite dangerous. I think that could be a conversation that we can have for quite a while.
Brett (51:46.859)
Yeah.
Yeah, if if their training sets are bi like the d the bu the biased data, if they were you know, if they're pulling from certain sources like the Reddits in the world that might be largely male, largely specific types of people, that's what they're being trained on and so they're gonna reflect some of the the you know it it's fe it feels like we're all in like, you know, middle school again.
Fern Potter (51:55.905)
Yeah.
it's idiots.
Fern Potter (52:03.038)
Exactly. Exactly.
Yeah, so it's like, it's that but then every mass adoption of that with no guardrails, no parental controls, no category exclusions, you know, there's no control over the access to that information. So the danger comes, I mean, we're going to have to go through this as individuals, like, you know, parents, aunts, uncles, whatever, go through how this is.
put slightly back in its box in the future. But I think that's one of the bigger questions I've gotten around the LLMs and this how it moves forward and what we can learn is the fact that what happened in social media is that they started an ad revenue model. So if you think about Meta, it didn't have ads to start with and it was pretty crappy. And then when it did, they were
the content was everywhere. then when it did have ads, it had crappy ads. But then as more and more consumers came and more and more brands invested their dollars, more and more safeguarding came into play. Now, I'm not saying it's a perfect environment, but you know, that generally is the trend as brands want to invest in that platform. And you create a ad network, essentially, you then have to answer to the money of the brand. And so the revenue starts to flow and then it stops if you're you know, the brand is then showing against.
Brett (52:55.487)
Yeah.
Fern Potter (53:20.778)
and savoury content, terrorism, whatever, all those awful parts of the internet that exist. We don't yet have that in LLMs. There's no... Is there going to be a moment, know, open AI is testing ads, there's talk of third party ad networks for the smaller LLMs that exist in different global markets.
Is that going to be something that then biases the information that we serve to serve the purpose of showing an ad for the brand to drive ad revenue? Is it something that's going to be a force for good because brands that can then actually help with consumer legislation on influencing content that shouldn't be served up?
I think we're going into a massive world of the unknown. think that's why I kind of relate a little bit back to like social media and that journey that it's taken in terms of ad revenue, because the brands just had such an influence on being able to protect the consumer. And none of that is happening at the moment in the GPT sense. When it comes to business, you've still got the same content that is surfaced.
Is that going to then influence if you've got an ad funded model, is that going to influence the model that then goes into your tools and technology? So I don't know, I kind of thinking about it in a broader sense of the consumer.
Rio (54:43.964)
When looking at how AI is being applied within companies, like you had this phrase, I liked it, thinking power, not people power. it was a bit provocative, I thought, and and kind of goes against that trend of we just gonna use AI to reduce headcount, right? like com like I mean, I guess that was a concern. Is that is like what do you mean by that? And and I know that a lot of the AI use cases initially were for
For streamlining things. I mean, there were a there were ads for AI companies that were saying like, don't hire a person anymore. And I just think that's terrible PR, right? And then there are the the prognostigations of doom from Dario and Sam Altman about all these jobs being wiped out, which which haven't happened, right? I see no no evidence they have. But thoughts on that, like what do you mean by that exactly?
Fern Potter (55:22.306)
That's it.
Brett (55:22.439)
Yeah.
Fern Potter (55:28.034)
Thank
Fern Potter (55:34.487)
Yeah, you know, there's no doubt that AI is going to change everything that we do in business as individuals. And I think there's what comes with change is a huge amount of opportunity and but managing that in a meaningful way. So I think we talked about it at the top. The training that consultancies are doing within brands is like how to use LLMs, etc.
What's the training that goes alongside that as to how your skill set needs to evolve as you start adopting more AI is a bit of a gap that we've seen within some companies that not just the change management and the practical application, but what does this mean for you? Critical thinking, know, strategic ideation, all these, that real thinking power skill sets that if you're operational, perhaps it's not something that...
Brett (56:04.317)
Yeah.
Fern Potter (56:28.076)
you've considered as core competencies before. And I think that will help rapidly change businesses because you're allowing people the space to actually start building out their own.
Brett (56:40.38)
And and doing higher value work, right? You know, and and like I think it comes down to if there's there's and we we certainly had some conversations about this, if there's like a data and information congruence, there's no gaps, right? Meaning people all have access to the same stuff, right? they can do higher value work, right, by accessing a system that allows them to to experiment, right? And to c and to ask questions and to you know, and and and we've we've had a lot of conversations about the sort of talk of the the hyphenated job.
description like people are people are the this the death of specialization we keep saying it ad nauseum but but I think that's sort of the direction that you're seeing within organizations. I mean or at least maybe not yet but I think that's the future direction of pe of people if you've if everybody's got equal access to information and knowledge based the based on that knowledge graph that's built within the walls and confines of an organization and governed appropriately, right? and it's headless and it gives out people equal access, allow them to to you know come up with solutions that may
not be within their domain of expertise, right? Like I I have a media strategy problem, but I you know, but I'm not a media strategist, but I still have access to some information and so I can get into a conversational interface and actually w l move down that path without having to actually be a specialist in that particular function.
Rio (57:56.711)
But you don't you know but I think it's interesting about that though, like before you jump in Fern is that I've noticed LLMs have gotten so smart. I mean, like it's I mean, you can ask it about physics, you can ask it about media planning, you can do anything it's gonna know, have access to more information, and be better versed in this than than almost anybody, right? And their and their reasoning abilities have gotten very incredible at the same time. So like we all have in theory, like everyone who
Willing can pay twenty dollars a month now has access to the most incredible tool in a lot of ways we've ever produced. But it hasn't been it hasn't changed things as dramatically as people have predicted, right? It's not like like having access to a two hundred IQ person. It's great, it's helpful. I'm way more productive, I'm enjoying my work more, it's definitely changing things. But it's not as though like
I remember those podcasts about a year ago where they were talking about a country of geniuses and a data center we don't have access to and how it was gonna overnight completely change industry, could change everything. It hasn't it hasn't had the it will have it's having a big impact, but it hasn't had the dramatic impact people predicted.
Brett (59:04.148)
Yeah. Well because because of where companies are in terms of their implementation of of this type of technology, right? Like unless you start with that sort of knowledge center that brings all this that connects to all the systems within an organization and gives people equal access to to that data and that information, then where are you starting from? This isn't about public LLMs. This isn't somebody using a public LLM. Yeah.
Fern Potter (59:13.558)
end.
Rio (59:25.254)
It it it also has to be filtered through people too, right? Like peop we're still the ones interacting with each other, doing our jobs. This is our world, right? So I think that it's made us better and it's automating things, but I think that humans still are the I didn't want to say limiting factor, I think that's the wrong word. I think humans are the conduit with which these tools will ultimately interface with our reality. And maybe in the future that'll change when you have i i i in machines interacting with machines on their own more often. But as of now, I don't see that changing.
Brett (59:54.857)
Well and it's it isn't it like humans, our organizational capacity, our ability to organize as as like as sort of like hives, as groups, is what is helped us dominate the entire planet, right? Like our ability to organize and work better together. But it sounds like a lot of the AI companies that I'm talking to in Fern, tell me if this this resonates, are are trying to bring that same kind of thinking.
into an organization to say, you know, y you've gotta be it's w you know, you've gotta be able to link everybody into one common place and that will the you know, cause the the it's greater than the sum of its parts, right? And and it's gotta s you know, an AI ha has to start there. It can't be just a fragmented bunch of people using it in different ways with no sort of underlying, you know, connectivity, right? Because then you're not really, you know, leveraging the power of the of the of the group.
Fern Potter (01:00:46.05)
Yeah. And what does good look like? What's a good output? Like if you haven't predefined as a business what a good output looks like, how then do you know what to put in and what to, you know, not just a knowledge base, but as an individual, if you haven't kind of sat and gone, let's use our human collective human intelligence to go, right, I know this looks like a really
Brett (01:00:47.261)
Right.
Fern Potter (01:01:09.16)
know, shithole comms plan for a specific brand product in Newmarket. This is what, and I know that because I've spent 20 odd years doing it. This is what good looks like. And this is the method and the steps that we need to do to get there. And this is a relationship you need to have with those people to do it.
Brett (01:01:16.617)
Yeah.
Fern Potter (01:01:25.198)
And these are the signals that will go in it. And therefore, what can AI help remove the friction from to your point then and democratize the ideation around that single platform. But if you don't know what good output looks like, everyone's kind of using stuff. that's Rio, you mentioned it before, that 95 % failure rate, I swear, is a lot of actually throwing stuff at the wall and see what sticks, but also not having clear defined quality, you know, idea or
construct of what a good quality output looks like from the start.
Brett (01:01:57.715)
Yeah.
Rio (01:01:58.097)
Well how are they defining failure? Failure is you use it and then you don't do anything with it. That's a I guess that's a failure, but that's not because of the technology. That's because you didn't have a plan and you didn't actually like have any thought about how you're gonna apply this or maybe maybe it was the wrong test. I don't know. So I think these numbers j
Fern Potter (01:02:03.266)
Well, aren't you?
Fern Potter (01:02:07.712)
Mmm.
But it comes back to that thinking power, doesn't it? That human power is like, what is what does the output need to be? Well, only you know that for your business and trial and tested and then using AI to get to that point. But I actually I was going to say on the failure rate, I kind of applaud it as well in a way that people are trying. It's very good education.
Brett (01:02:28.945)
Yeah. So so
So how do you guys think about the you know, 'cause I've talked to a few companies that talk about the you know, sort of like this notion of of building blocks or primitives, right? And then and then specific skills like jobs to be done, not skills in the claude sense, but jobs to be done that plug into the building blocks that are basically like compute reasoning power, and then you have a very specific job to be done, you know, build PowerPoint presentation based on, you know, an entire design system and every presentation this company's ever built, right? And so you've got the the primitive has got all the design systems in in the build
Blocks have got kind of all that core learning data and information. Let's say it's every presentation the organization's ever made that's stored, that's all been compiled and brought into. And then the skill is the ability to roll out white papers, one-sheets, PowerPoint presentations, all in a consistent way, with a consistent voice and a consistent format. That's a very practical.
ex you know, way of doing this, of leveraging like a co you know is that how you guys are thinking about architecting it, where you sort of build like a compute reasoning system, a building block, and then you and then you start to leverage start to build skills specific jobs to be done on top of that? I mean, how are you thinking of it?
Fern Potter (01:03:37.903)
Yeah, essentially, there's these, these kind of, as I we call them engines, I guess, in the or not, it would be skills would then set up and have different jobs and commands, and they can be sequential, they can be just modular as well. So from one stage to the next, you kind of work through a whole business intelligence system that gives people what they need in their hands based off the knowledge base that's there. So, as you said, whether it be
Brett (01:03:51.901)
Yeah.
Brett (01:03:59.816)
Yeah.
Fern Potter (01:04:03.69)
a common strategy, whether it will be a deep research analysis for a specific product, but it not just based on where the market is now, where the market is going, but where you've also been, but also
Brett (01:04:12.957)
Yeah. And w and what is this engine? What does the engine look like? I think the engine is analogous to the the primitive or the or the building block, right?
Fern Potter (01:04:19.348)
It is, yeah. So essentially it's agents, it's APIs, it's knowledge base, it's everything that we can build into the intellectual property of that single engine that belongs to the business. we don't look at AI in isolation, we look at it as part of the bigger system to make sure that it's fully kind of autonomous within the business systems of a company.
Brett (01:04:34.003)
Yeah.
Fern Potter (01:04:44.182)
and it's got different tooling and practical applications that businesses can use.
Brett (01:04:48.583)
Yeah. And is there one engine or are there multiple engines built for different sort of com yeah.
Fern Potter (01:04:52.174)
Yeah, different functionalities, multiple engines. Each one is custom made to each of the businesses that we work with. They might do the same as you can imagine to make sure that it's doing the right job for that business need.
Brett (01:04:59.347)
Yeah.
Brett (01:05:05.831)
Yep. Yep. And then and then I'm sorry.
Rio (01:05:07.434)
So looking at looking at looking at like agencies, consultancies, and different professional services orgs that are not only advising their clients how to adopt AI, but being impacted by AI themselves. How do you see these business models being impacted as labor changes, as roles and skills change, as I mean, even the having things for junior people to do change a little bit and and
And what senior people do and how you build clients. Like how is that are you seeing those impacts on your clients today? And what are they?
Fern Potter (01:05:42.583)
Yeah, it's such a great question, think depending on the different clients that we work with and where they're at, that change is happening. the commercial model, it's interesting. The commercial model will have to change, but I was thinking in my head as you asked me the question. The barriers...
to entry that we've seen so far, like the business frictions have been more on the development side of business than they have been anywhere else. So in those kind of Dev and tech teams, yeah, into engineering. I think, you know, we'll come across a business and we're like, yeah, this is great, but we've got a 12 month DevCube locked in. And you kind of...
Rio (01:06:15.259)
Engineering.
Brett (01:06:24.371)
Yeah.
Fern Potter (01:06:27.446)
I know, I think I anticipated obviously wrongly that that would be where the agile part of the business is that would have that focus on AI and it's not there.
But I've seen in other businesses, well, it's day rates, it's et cetera. know, is projects that have been committed to and you get that. But I think that commercial rigidity will have to change. And we look at paying, you know, there'll be commercial models that are focused on goals, on deliverables, on acceleration of product that is shipped, you know, MVP versus XYZ features adopted. So I think that will probably change.
And I'm sure other world commercial rigidity has always been a bit of a hindrance even when it comes to media, right? If someone gets paid 10 % on a media spend and you want to reduce it because you're going to look at high quality media environments, they're not going to be happy. like all these things, the commercial structures are going to have to have to differ. We we've seen, like I said, we're a SaaS business. So build the product and then the ongoing iterations for licensing and maintenance. But often.
Brett (01:07:30.905)
i and you guys charges it well, I'm surprised you're using the word sass. I thought that was a dirty word nowadays.
Fern Potter (01:07:34.924)
Well, was going to... Yeah, well, we kind of see it as like services as software, right? You're flipping the switch a lot. It's very... Yeah, it's... Yeah, it's got very different connotations, hasn't it, these days, actually? So...
Brett (01:07:39.591)
Yeah. services as software as opposed to software. Yeah, that's that's or solutions as a yeah, it's yeah.
Yeah.
Rio (01:07:50.333)
Well, service is making a big is having a big rebound. Yeah. I mean, like all of the the forward deployed engineer, which is really a consultant, which is a services org. I mean, it makes much more sense for these AI companies to have a serv have a services component, because the margins are actually more similar to the AI component, which is interesting as a as opposed to traditional SaaS, which had crazy margins. So these SaaS businesses never wanted a service component. I think it's actually interesting. And I think there's also the argument that you can't deploy AI without services. It's just not happening.
Brett (01:07:57.245)
Yeah.
Fern Potter (01:07:58.702)
Bye.
Brett (01:08:13.778)
Yep.
Brett (01:08:17.905)
Yep. Yeah, yeah.
Fern Potter (01:08:18.686)
car yeah exactly
Brett (01:08:19.974)
Yeah, and it's it's sort of a it's a challenging sort of and I'm dealing with this with the client is a it's a challenging positioning play of like is tech enabled services does that sound like the tech is first or the people are first, right? Which comes first, who's leading, who's following, right? And you you run into some of these challenges because you because and I and I think to your point, Rio, is that the the managed services model of like we're doing this but we're doing it one one tenth one tenth or one one hundredth the time, at a level of quality that you couldn't have gotten in
Fern Potter (01:08:30.734)
See ya.
Brett (01:08:50.139)
in in the past with fewer people because they have this this decisioning engine basically behind them which is which is tech.
Rio (01:08:58.822)
it's it's it's incredible. So, Brett, I'll give you an example. The other day a client reached out and they wanted me to, they're deprecating a piece of software and they said, I just want you to come, can you just give us a quick point of view? And like, look, I I've it's a type of software I've worked with for many years. I know this space well, but to think about to actually think about it, give them a list of quick list of vendors to look at, give them four or five different like operating models or deployment like architectures they should could consider, not giving them recommendation, but just telling them guiding their way.
I crank that out in two hours. That would have taken me a week before, right? And I'm someone who knows this space well. If junior people would have taken to a team of two or three a couple of weeks come with a point of view like that. I'm just thinking, and I I said I don't I'm not charging you for it. Don't worry about it. Take this and then if you want to bring us me back to to do a paid engagement with a couple other people, we can do it. We can actually execute this and give you the rigor and do dealers you're gonna need. But the fact I could create that up right in a couple hours is absolutely mind blowing.
Brett (01:09:38.706)
Yeah.
Fern Potter (01:09:57.197)
The fact that your experience meant that you understood the problem and then could crank that out is something that, you know, yeah.
Rio (01:09:57.233)
Right. I mean
Rio (01:10:01.616)
Yeah, a junior person couldn't have done it for me. There's no way. I I had to know like I had to guide it what to look at. I I get I I had to feed it like things I had done before that were similar. I I think I talked to it for like ten minutes of thinking through things and I went through a bunch of iterations. But the fact I could get through something that complex in a in a couple hours, like again, I've been in consulting for fifteen years. That would have taken me a a a minimum a week before with
Brett (01:10:01.757)
Yeah.
Fern Potter (01:10:06.871)
Yeah.
Fern Potter (01:10:13.614)
.
Fern Potter (01:10:19.47)
Yeah, it's incredible.
Rio (01:10:27.184)
Not maybe not full time, but working on the side with another person helping me and just it's incredible. I said to the client, I'm not gonna it's Jason, Well I d we can't pay you for this. I said, Don't worry about it.
Brett (01:10:34.576)
Yeah, you you you Yeah, and you've much you've moved much closer to actually being it's almost like a product leadership role, right? Of the days of past, right? Hook you up with an engineer and you could probably you've basically built out what could be or at least could evolve to a product spec, right? And you're not a traditional product person. I I found myself that was one of the reasons why I jumped into startup space, Fern was like I I wanna like I I take my creativity to the product world. I don't have to be the full stack developer engineer.
Fern Potter (01:11:01.356)
Yeah.
Brett (01:11:04.124)
But I can be the product leader to come up with the idea and then and then because of democratization of sort of information access, because it's so easily accessible, you can get a ton of intel of what are the best ways to structure this, to spec this out, to communicate in ways that the builders can actually build it, right? And and I've I've just, you know, but been able to that would have taken years of experience on the job to develop. And it's just stuff I've been around this stuff for so long, now I can actually bridge that gap in like much less time.
Right. and I I think that's one of the values of AI is it just allows it does democratize a little bit in that in that respect.
Fern Potter (01:11:34.882)
Yeah.
Fern Potter (01:11:39.222)
Yeah, that's the thinking power, You your experience in the creativity that you can bring and then be able to accelerate it into just places that you wouldn't have been able to before so quickly.
Brett (01:11:51.388)
Yeah.
Rio (01:11:52.562)
Should we move to quick hits?
Brett (01:11:54.001)
Yeah, totally. Yeah, I I I think Fern needs to eat. It's it's getting late over there. She's Yeah, she's got a cold pint waiting for her. you more of a cocktail type. More of a cocktail type, not a cold Yeah. Or a s or a slightly room temperature pint, you know, you know the old anyways. the bit the bitter, right? They call it the bitter, right?
Rio (01:11:57.459)
Well we're standing between her and happy hour and her probably hitting a pub or something, so
Fern Potter (01:11:57.551)
Fern Potter (01:12:08.01)
Absolutely, yeah, you know. I'll take a cocktail, yeah.
Rio (01:12:16.636)
Bitters, yeah.
Fern Potter (01:12:17.966)
I just didn't... Yeah, I don't... I'm not, I'm not...
Rio (01:12:20.604)
Yeah. I'm a I'm a fan.
Brett (01:12:23.176)
You're you're a fan?
Rio (01:12:25.106)
Bitters. I love
Brett (01:12:26.16)
I like I like it better. It's it's a it's it is a little strange that it's room temperature, but but
Rio (01:12:30.812)
But no, but it but they are hard to harder hard to find. It's funny. So I I was invited to before we go to Quick Hits, I was invited to like a c to visit a client in Manchester. And and I hadn't been in Manchester in a long time, so I go to visit them and they were and it's funny, they c they're all waiting for me, like, the American guys in town. Let's we're gonna take you to like the best like b place to get beer, because y you know, I told them right. No, but they took me to brew dog. And I was like, brew dog, this is like the
Fern Potter (01:12:50.195)
Yeah.
Brett (01:12:52.474)
It's always some old school pub, right?
Fern Potter (01:12:54.623)
All it is.
no!
Rio (01:12:58.63)
Th this is like double IPAs and like just the cry like the stuff you get here. I can go to like a hundred places in Denver and get like better versions of that, right? Like tr like like hazy IPAs. Yes, yes, yeah. I want bitters like hand pulled, you know, like where they serve meat pies, right?
Fern Potter (01:13:03.99)
my god.
Brett (01:13:07.194)
Yeah. I want a dirty old pub with people that have dirty teeth, right? Yeah, they're they're pumping that that that that that cat that cake. Yeah, yeah, there's peas and potatoes. I mean I want Yeah. That's the English experience.
Fern Potter (01:13:07.918)
I love that they were...
Fern Potter (01:13:13.5)
Fern Potter (01:13:23.894)
That's the name, that's what you need. That's a great night out.
Rio (01:13:26.194)
What you need. I was, I was, I told it. I was I'm not I did not envision broodog. Can we go somewhere else? And they brought me somewhere.
Brett (01:13:31.496)
This is what yeah. This is like post like a postmodern nightmare in England. You're like, I want traditional England. anyways.
Fern Potter (01:13:32.514)
Did you? love you. I had to. I had a couple of weeks ago did the same and one of the guys is like, you've got to take them to Wetherspoons. Do you know the Wetherspoons over in the UK? It's like it's just cheap, awful pub that's open from really like really early in the morning. And it's kind of it's I don't know. I think it feels a bit like a social experiment more than a pub. But so.
Brett (01:13:53.853)
Yeah.
Brett (01:13:59.087)
Yeah.
Fern Potter (01:14:00.047)
I won't recommend you to go. However, there are some beautiful pubs that serve cool pints. Just want to make sure that I'm still in the British dream.
Brett (01:14:06.437)
Yeah.
Rio (01:14:08.956)
Well well neck well sometime you'll have to take us at a a driving tour of the Shire and sh show us some good pubs. That would be that that would be amazing. Be amazing. All right, so
Fern Potter (01:14:15.212)
I'm digging it, you're coming. You're coming over. I'll organize a tour, some historical events.
Brett (01:14:17.893)
Yeah, yeah.
Alright. Yeah, all right, definitely, definitely. So quick quick hit. So so most overhyped AI use case that you're seeing.
Fern Potter (01:14:30.764)
And so the one that really annoys me is people using it for their LinkedIn posts.
Rio (01:14:35.78)
LinkedIn LinkedIn s AI slop, right? The worst, yeah.
Fern Potter (01:14:39.32)
so bad. It's so obvious. It's just so obvious. I just like not really know what to do. Sorry.
Brett (01:14:40.283)
Well if you train it well enough.
Yeah.
Rio (01:14:45.49)
Okay. What's one?
Fairness, one task in your life that you've either already or would happily delegate to AI.
Fern Potter (01:14:55.18)
I mean, one that I would love to delegate is my tax return, but I don't think we're there yet from a government perspective. But definitely the thing I do love is just all the meeting recordings, all that qualitative data that you can get in a single place right now and be able to look back on, but also have a singular record on that feeds your own knowledge base. I think it's fantastic.
Brett (01:15:17.659)
Yeah. And there's a lot of cool use cases in AI around that, around not just not just creating a folder with meeting notes that have been, you know, carefully curated by your AI bot, right? and and you know, and then going it but something where it just it absorbs all of that across all of your clients and then allows you to to conversationally interface. We're actually building something like that internally within within our org, yeah. Which is I mean
Fern Potter (01:15:25.122)
Excellent.
Fern Potter (01:15:33.62)
And you've
Fern Potter (01:15:38.126)
I like you. Oh,
Brett (01:15:41.466)
Leveraging a partner to do it, but but point being is that that I just think there's it's again it's like how do you leverage that intelligence of all of your meetings with a client over time as opposed to having to c ask the L LM to go and review all the files all over again, right? It sort of builds a knowledge graph around that particular client and all of the interactions you've had over time with them. And it doesn't waste all the tokens that it would take to just keep doing that analysis every time you want to do an evaluation, right?
Fern Potter (01:15:55.552)
exact.
Fern Potter (01:16:01.966)
Yeah.
Fern Potter (01:16:05.822)
Exactly, every time you have to run it. yeah, and even like, know, you've got these, what do I use, like, you talk to yourself, those dictaphone and stuff, as you'll go around and be able to put that in together with your ideation. I just think it's brilliant that it's just always with you.
Brett (01:16:15.524)
Yeah.
Brett (01:16:21.669)
Yeah, yeah, for sure. So, one well actually what what's the bigger AI constraint? Technology, data or organizational behavior?
Fern Potter (01:16:35.374)
I say tricky. got to can I answer two things? One I'd say is psychological constraint. It's change management. We always kind of had a bit of a joke that would have our next, you know, senior high might be somewhere in like psychologist space. But just in the experience that we've had so far, and I think we'll get there. I think it's just nervousness and trust issues. And I think that's absolutely fine. Like, that's the way it's going to be.
Brett (01:16:40.229)
Yeah. You can answer.
Brett (01:16:45.862)
Yeah.
Rio (01:16:45.938)
Interesting.
Fern Potter (01:17:05.208)
The other thing we mentioned at the top of conversation that I'm really passionate about is just solving for the gender bias that exists. And that's a huge constraint because it's not just gender bias. There's so many biases that exist within the LLMs. If you are building on top of those in a way that is taking that information and making decisions for you, especially when we come to that full end-to-end autonomy, it's actually widening the gap of...
Brett (01:17:31.674)
Yeah. Yeah, there's a lot of talk about like AI ethnicists. We talked to Sunni, right? The the chief AI officer of Denver that talked about how they brought in a Buddhist monk.
Fern Potter (01:17:36.558)
Thanks for coming.
Rio (01:17:41.298)
Suma, yeah.
Brett (01:17:42.352)
S s Suna, did I say that like I screwed up her name. See, I assume I totally screwed her name. but like but ethn yeah, having like it's this level of sort of morality. And again, it's thor theoretical, but how how do actually apply that? to to the learn to the learning engine so that they can apply some level of it's like a it's like a governance layer, right? It's human you know, morality or ethnicity or whatever or morality governance, right? It's in a sense.
Fern Potter (01:17:42.53)
I want.
Rio (01:17:44.38)
Showma.
Fern Potter (01:17:47.01)
Yeah, I see.
Fern Potter (01:17:55.628)
Yeah, how do you feel this?
Yeah.
Fern Potter (01:18:08.878)
It's in the development, in the build, it's in the information, it's in how you then translate the information out the models and make sure the information is complete. It's about spotting bias and then being able to complete it. I think there's a lot of movements as well as research on how we can be better, basically.
Brett (01:18:27.065)
Yeah. Crazy complex but important.
Rio (01:18:30.204)
Fern, what's one lesson from programmatic advertising or ad tech that the AI industry should learn immediately?
Fern Potter (01:18:39.726)
I mentioned it before and it's something that again I feel quite strongly about is the way in which you build with simplicity. There's a lot of complexity that exists within AI. shouldn't have to be that's off putting. think the...
the way in which you operate and how you work as a business with other businesses should be in the most simplest form possible to get people started on that journey to what AI can help them do for growth. Like no use case is silly. Like there's nothing that you shouldn't, you know, I've had this, you in these WhatsApp groups and stuff, that's pretty basic. Well, good, it should be, it could be basic because that basic can be really fundamental change to a business and the way they operate. So.
Brett (01:19:17.157)
Yeah.
Fern Potter (01:19:33.006)
I think we create a load of programmatic complexity and I don't think we should do that when it comes to AI.
Brett (01:19:40.496)
Yeah, less is more, start simple and tell Rio to s to turn off his his vibrating phone. Right? From this episode. That's what the dude has to serve us for. It won't vibrate. I kid you. I d I don't know if our audience can hear that, but yeah, so finish this sentence.
Fern Potter (01:19:48.416)
So fun!
Fern Potter (01:19:56.078)
What could this sound like for the whole thing?
Rio (01:20:00.079)
Who well did you did you hear earlier my wife actually you know you can do do the thing where you ping your phone and like it makes the noise.
Fern Potter (01:20:07.896)
We just walk?
Rio (01:20:08.018)
She did that, she No, you can ping your phone from your watch, right? If you've lost it. And like I guess she left her phone right over there. I was like, no. but it was her phone. Yeah, it was her phone. Yeah. So she's trying to find it. Yeah. So and
Brett (01:20:08.032)
you can ping you you can ping somebody and
Okay.
Fern Potter (01:20:17.294)
I'm sorry.
Brett (01:20:18.459)
is is it her phone that's vibrating? She's trying to find her phone. my god, that somebody are like, This isn't even my phone. I don't even I can't even turn this off. anyways, so
Rio (01:20:28.4)
A anyway, so but the fern, this has been super fun. This has been great. This has been and like we are standing between you and your your your your cocktail of choice. so have a great weekend. This was like fun hanging and talking to you and it was it was as it was as enjoyable as we thought it would be, right, Brett?
Fern Potter (01:20:32.627)
it has.
Brett (01:20:33.563)
It's been
Brett (01:20:44.849)
Total a great Friday afternoon and a Friday s early evening for you. well well hey, thanks again. We'll we'll certainly have you on again and when you're in New York City, definitely give a shout. Same same with London. well you are a little further from London, but you know.
Fern Potter (01:20:45.566)
brilliant.
Fern Potter (01:20:57.888)
I work. But you've got to come back because we've got to dispel all the myths that you've created around the Shires.
Brett (01:21:03.715)
Yes. I yeah, and I need I need to do a proper travel around multiple parts of of like I've been to Bristol and Cambridge and London, but there's parts of like I've never been to Manchester, I've never been to Liverpool, Birmingham just for the the sake that it's where Black Sabbath is from. Yeah.
Rio (01:21:04.09)
Yes, yes we need to do that.
Fern Potter (01:21:17.327)
you can count them.
Rio (01:21:22.02)
I've heard York is quite nice. Never been.
Fern Potter (01:21:23.778)
You've got to come. York's beautiful. Yeah, very historic. We'll do a tour and then we could do the podcast like in the car. I can imagine you have a carpool like tour of the UK. Exactly. No, thanks guys for having me. I really appreciate it. It's so nice to speak to you. And thanks for your time and great conversation.
Brett (01:21:30.151)
Yeah. The English countryside. All right.
Rio (01:21:31.88)
I love it. We'll get a Starlink and we can do that.
Brett (01:21:43.001)
Yep. And for everybody that made it this far, w wsignal and noise dot AI, you can find us on TikTok, Instagram, Apple Podcast, Spotify, YouTube, or just visit our website. And we'll see you next time. Thanks everybody.
Rio (01:21:53.254)
You name it.
Fern Potter (01:21:58.031)
Thanks, bye.
Comments