AI Doesn’t Fix Data Problems — It Amplifies Them: Acxiom's Crystal Wallace on Governance, Agentic Workflows & the End of “Monolithic SaaS”
- Mar 9
- 49 min read
In this episode of Signal & Noise, Rio and Brett sit down with Crystal Santos Wallace — enterprise operator, data leader, and longtime architect of modern marketing infrastructure across global holdcos — for a candid conversation about what AI is actually doing inside agencies.
Crystal has worked across the major networks and now operates at the intersection of data, identity, and transformation at Omnicom and Acxiom. She’s seen data evolve from passive record-keeping to operational truth — and now into fuel for agentic systems.Her thesis is simple but sharp: AI doesn’t fix broken data. It makes broken data louder.
What We Cover: From Platformization to Agentic Workflows:For the last decade, marketing technology has centered on “monolithic SaaS” platforms promising integration and control. Crystal argues we’re entering a new phase — not the death of platforms, but their transformation. The future isn’t fewer systems; it’s orchestrated systems, powered by agents and governed by design.
Governance Is Not Optional:As AI accelerates, governance becomes existential. From model access and data leakage risks to inappropriate automated outputs, Crystal makes the case that compliance, security, and policy must evolve alongside innovation — not trail behind it. Human-in-the-loop isn’t a philosophical preference; it’s a risk mitigation strategy.
Synthetic Audiences & Continuous Learning:We explore the rise of synthetic focus groups, digital twins, and always-on modeling. Used correctly, these tools compress the research cycle and create test-and-learn loops at scale. Used carelessly, they amplify bias. The difference? Data quality and disciplined inputs.
The Changing Commercial Model:As AI reshapes workflows, the traditional FTE model inside agencies comes under pressure. Crystal speculates about usage-based, tokenized, and outcome-oriented commercial structures — and why the winners will balance technical literacy with strategic altitude.
The Future CMO: Tomorrow’s CMO isn’t just a brand steward. They’re an orchestrator of systems, a translator between data science and business strategy, and a leader who understands enough about AI to direct it — without being consumed by it.
This is a conversation about scale, identity, governance, and the pace of change. It’s also a reminder that technology doesn’t absolve us of responsibility. It magnifies it.
If you’re navigating AI transformation inside an agency, a brand, or a data organization — this one’s for you.🎙️ Subscribe to Signal & Noise wherever you listen.
Read full transcript bellow:
Brett House (00:01)
Hey everybody, welcome back to Signaling Noise, a no BS media franchise decoding data, tech and AI and the marketing transformations that actually matter. Be sure to subscribe today whenever you can, ⁓ wherever you're listening I should say, or watch today's show on Spotify, Apple Podcasts or YouTube. Today we're joined by Crystal Wallace, someone that I've had the pleasure to get to know over the last few years. She was...
She joined our Brave New Worlds show at TransUnion and New Star. ⁓ And thank you for joining the podcast today, Crystal. ⁓ Thrilled to have you. She's an enterprise operator. His career has been shaped by moments where data moved from passive record keeping, which goes way back. It's like when Rio and I talk about how far we go back in the advertising world. It's like pre-internet to operational truth.
Crystal Wallace (00:53)
Ha
Brett House (00:56)
And she's worked inside and alongside organizations like Omnicom, IPG, Publisys, Indentsu. You're currently heading operations at Omnicom, I guess, officially. And Axiom, yeah? And you're gonna share, I think, a really good operator perspective for our listeners on how data AI and compute are reshaping how organizations function, especially agencies.
And welcome to the show Crystal, thrilled to have you.
Crystal Wallace (01:27)
Thank you so much for having me. It feels so good to be part of this brain trust. Brett, you're brilliant and hopefully will break me up my shell. This is not my natural format. I'm an intro extrovert, but yes, I've been across all whole coast holding companies, but data has been an ethos. Even as a child, my dad was computer programmer for IBM.
So I was playing on BBS boards, if you remember those back in the day. And then as Brett mentioned, of course, I worked for Emory in the postgraduate fellowship program transferring, you know, hard medical records into e-records. And I've just seen a lot of, you know, when you get to this inflection point where data really becomes, you know, a currency and an expression of intelligence. And I think we're obviously in that phase now. So, ⁓ but there's a lot of noise. Yes, for sure.
Brett House (02:30)
Yeah, Rio and I joke about the fact that we both are old enough to know when internet used to be delivered on disks that arrived in the mail. Right, the old AOL days, right? Sign up for an email address today.
Crystal Wallace (02:41)
I would like to.
Rio (02:43)
Or even before that, remember the modems you'd plug back into the back of your computer, the Commodore 64, right? I that was where I learned how to program basic. It's funny. I guess it does date all of us, doesn't it?
Crystal Wallace (02:46)
Yeah Yep. we're gonna all. But. ⁓
Rio (02:56)
Well, Crystal, we are thrilled to have you on here, Crystal. We were excited for this episode for a bunch of reasons. mean, you you can go break away back. We just met recently and now that we're colleagues officially, the merger happened. I'm really thrilled to get to know you better and have you on here. for those who don't know, which is probably no one at this point who listens to our podcast, IPG and Omnicom did have a merger recently in Axiom division, which is I was so excited to work with people that actually are so many good friends over there from over the years. So it's, I think it's a great thing. It is a lot of, a lot of like synergy we're going to unlock over the next few years. I think it's already starting, but having you on was great for a bunch of reasons, ⁓ include, especially your resume. mean, you've like, you've done the tour to hold co, right? I mean, you've, literally been everywhere. So you have some amazing experiences and I, and like, we're thrilled to share with our viewers. You're.
Crystal Wallace (03:43)
Hahaha
Rio (03:50)
like your points of views and what's really going on in the, in the worlds of Martech and Atttech and how that's impacting marketing and how that's impacting the HoldCo model, right? So you think about the last 10 years or so, there's been Martech and Atttech ecosystems really built around these big, you might hear this term platformization, right? These large monolithic SaaS platforms, integrations, they promise integration control and centralizing workflows. And I that's been an emphasis over the, the past bunch of years, you've seen the rise of lot of these platforms. And recently, the intersection of that and AI, and wow, what does this mean? Is this going to change the way we operate? Is this going to change the way media is bought, sold, planned, optimized, et cetera?
What is the impact? Where are the investments going? How is it impacting the way we work? And how is it impacting the way our clients expect us to work? So think there's a lot of really cool discussions that come out of this. with that said, welcome, Crystal. And ⁓ Brett, I think you had a couple of questions you wanted to start off with,
Brett House (04:46)
Yeah, yeah, well, I you saw it in the news ⁓ recently. mean, as of today, although this episode will post a little bit, post this news, but with Frontier from OpenAI and Claude Cote and the cowork announcements from Anthropic, not to mention those really funny Super Bowl commercials that Anthropic put on. I know a lot of people are flipping out about anachronisms, like Eric Sufer was like, this is an anachronistic take on the media industry. ⁓ But.
Crystal Wallace (05:05)
Haha.
Rio (05:05)
Yeah, those are coming out Sunday. I call this sanctimonious.
Brett House (05:15)
Yeah, sanctimonious, I must
Rio (05:15)
I will say that.
Brett House (05:17)
admit as Sam Altman admitted, I laughed. I laughed my ass off with the woman, with the mother conversation. That was really funny. But these announcements ⁓ seem to be ⁓ really impacting the investor community and how they're seeing monolithic SaaS, right? And they're in a sense, there might be a panic mode right now.
Crystal Wallace (05:18)
It was awesome though, come on.
Rio (05:23)
They're pretty good.
Crystal Wallace (05:25)
Yeah.
Brett House (05:40)
They're panicking because they're seeing these impacts of sort long horizon, agentic capabilities that can do almost replace a person with repeatable, functional, orchestrated tasks, not just sort of discrete single tasks, right? And it's starting to, people are starting to look at the legal industry, these other industries that are supported by SaaS or otherwise. I mean, what do you see that impact? And do you think that the market is this question was not the first question I listed, but it was sort top of mind. ⁓ Do you think that the market is overreacting? Or do you think that really this is going to play a huge role in sort of the future of work and the future of technology, within agencies or otherwise?
Crystal Wallace (06:26)
100%. ⁓So it's powerful, it's coming, and it's coming soon. I think there's an over indexing right now of how much it can do autonomously, right? We need human in the loop. And for me, I'm really excited about the ability to elevate and control more kind of mind space and think, ⁓ know, elevate and do more because right now we live in a very kind of reactive state. That's both, you know, just from a personal standpoint, but even the workforce, everything is, you know, always on instant. think maybe it's some byproduct of COVID, but ⁓ I think there's over indexing on you know, ⁓ the workforce is going to reduce. Instead, we should get a smarter workforce doing higher level things with intelligence that it's feeding. And we can give AI, MCP, your agent's forms, your agent societies, the tactical things that kind of like bear us down is kind of what I'm thinking. We do have to change the model.
Brett House (07:51)
Yeah, raises the value. kind of forces people, especially new employees, to sort of quickly move up the value chain, right, in terms of the strategic value they're delivering. Right?
Rio (08:03)
They almost have to.
Crystal Wallace (08:04)
That's what it is. And if you ever, if I am a master prop engineer and so are you all. You should, I mean, it's like in the early days of search, if you don't know how to search the right way, you're not going to get the results that you want. That's the same thing right now with these agents. It's not, they're still, mean, in this company, I say that today and then like tomorrow there'll be a release where it's doing something else crazy. But, ⁓
Brett House (08:13)
Hahaha
Crystal Wallace (08:33)
You still need to know how to prompt deductive reasoning. This is where common sense comes in really good order of operations. is still kind of a very much, know, it is a system. And so if you understand kind of system dynamics and design strategy, then it can do some really powerful, amazing things. But human in the loop right now is still paramount. I don't think that
Brett House (09:02)
Agreed.
Crystal Wallace (09:03)
People are just starting to tinker, the general population, the general work population. And I'm pushing all of my staff, my team, everyone champion, like go hit, you know, GBT or co-pilot or Anthropic, know, perplexity and see what happens. Some are good at, you know, almost like...robotic processing automation and they can watch and then take and automate. Some are not good at that. Some are better at writing. Some are better at coding. And there's a whole ecosystem and it's not all the same. And so understanding what model to use for what. And of course you have the MCP players creating kind of that logic and reasoning. But guess what? That comes with a cost, a toll.
So I think we're still in early stages. mean, the possibilities are endless, are gonna be amazing, but it's gonna be a while, I think, until you have, what is it, AGI and things of that nature where humans can just sit at home and do nothing, make the bots work for them. So we're not there yet.
Rio (10:21)
Yeah, yeah, yeah, maybe one day, right? mean, yeah, I mean, it's, don't know. There've been predictions about AGI being two years away or 30 years away about it. And I agree. I don't really think it, I think it doesn't matter, right? The fact is these tools are very useful.
I think you said a couple of really important things there that I want to call out and I agree with. I number one is about knowing how to prompt engineer. I mean, look, it's not rocket science to be able to prompt engineer, but I have noticed more senior people with more experience and better context and better subject matter expertise are able to get a lot more value from the large language models, from the AI tools. It's really interesting when we're comparing the outputs from them, from different people, I compare what I'm getting, I'm, let's say, working on a project with junior people, it's much better. Honestly, I can get a much better outline than someone who doesn't have the right context to put it in the first place. So I find that's very interesting that more senior people actually, your value's almost amplified by it, right? I think that's number one. I think the other point you made that's really good, and then I do have a question after this, the point was about them all being different.
They're evolving, they're evolving all the time, they're always pushing updates, many people don't realize Anthropic is mostly used B2B, right? It's used because it's used for writing code, it's amazing at that. Granted, the new OpenAI release is apparently pretty good, I haven't charted it out yet, but then OpenAI is mainly used for writing, so they all have their different things they're good at, so I think that's a great point, but I guess the question I had for you is looking at the workforce, like how are you seeing our clients?
I mean, are clients expecting the things to be done quicker? Are they expecting the things to be done more efficiently? Are they asking for discounts as a result of AI? I'm curious what you're seeing kind of out in the field.
Crystal Wallace (12:01)
Well, let me, I'm going to ask the last question. Clients are always asking for discounts. They have sweet. So like, I don't know. mean, of course they expect to do, you know, more, less and smart, but I'm not actually seeing that as a push. There's, they are absolutely expecting, um, to see how the agency leverages AI on their behalf to be.
Rio (12:07)
Okay, fair enough. We've trained them. ⁓
Brett House (12:07)
This is gonna be more excuse this is gonna be more of an excuse to ask for this good
Crystal Wallace (12:30)
to drive modern marketing and all the way throughout. you can use it for contracts, you can use it for synthetic personalities, you can use it to optimize the brief. So what are the aspects? And they wanna see it show up in that end-to-end workflow in a way that is privacy safe. The other thing that I am seeing as well is they wanna be able to connect ⁓ with agents on the agency side. want help building their agents and transforming how they leverage and taking all these models. ⁓ Can they do more in terms of modeling faster in real time? So what can you do to really enhance that? But I don't see it as cost, pressure, or constriction. ⁓ Yet, I think we're already, there are other factors that they've been trying to negotiate in terms of cost.
Brett House (13:40)
Yeah, that's something that we talked about at advertising week actually this year. And I brought this up before, but just to that point of cost pressure is I think it is going to put cost pressure on the FTE model because of the compounding effect, especially to your point Rio, of the more senior people that I think the critical thing with the seniority of people is systems thinking.
Brett House (14:05)
It's the ability to systematically, to develop and orchestrate programs, which is very systems thinking oriented. And you brought that point up as well, Crystal, because that's what allows you to orchestrate an agentic workflow or a combination of agents to do something that's more than just ⁓ a discrete task. It might be a set of tasks combined to a bigger thing, you know, a playbook or strategic play or whatever. So I think that's,
Rio (14:10)
I like that.
Crystal Wallace (14:11)
Yeah. That's right. That's why I say the model is wrong. So the FTE for sure like that. However, I think that there's a different way commercial model. And right now they're overpaying for platforms and not realizing the full value of them. that's when you need consultants. You want us to use AI, it comes at a price.
Brett House (14:55)
Yeah, that's yeah 100 %
Crystal Wallace (15:02)
Like it is not just, mean, if you have pro or you have, you you know how much you're paying for those token and that usage fees. So again, it's a different model, not saying, you know, and I think you can, think smart agencies, you get ahead of this. not, it is understanding kind of what things you can start to shift. You need to elevate, you need to invest in.
Rio (15:11)
It's not cheap, yeah.
Crystal Wallace (15:29)
⁓ to drive those capabilities of the future. So it's system integrators, it's, know, the data science, it's really how do I make everyone a strategist? And generalists probably at the same time.
Rio (15:43)
You mentioned a token based model earlier. I love that. It's the first time I've heard that ⁓ for the agency. Like I love if you can maybe elaborate on that a little bit. how do you, what do you think that looks like? How does that change the way we function? We operate, service clients.
Crystal Wallace (15:57)
It's totally half-baked, this is probably going to be a bad answer, but I remember when I was at another whole company, this was probably ahead of its time, but we were in a platform-based world. Now you're hearing that all the time, platform world. ⁓ And that was actually ⁓ based on all the platforms that we were having that we had to pay for in order to service the business we always serviced. Now I see like where are the cost in terms of the supply chain? Where are the increasing costs that I need to look at and plan for? And that largely is around AI and ⁓ usage, whether it's in the cloud, right? ⁓ You're paying usage-based fees in the cloud tokens, cetera. And so that's where my token concept comes from. I used to think maybe you were going to be a crypto or bit, but you see what happened to Bitcoin yesterday.
Rio (17:02)
Yeah, it was an annihilation. It went up today, but it was pretty bad.
Brett House (17:02)
Yeah
Crystal Wallace (17:06)
I had a second, I was like, ⁓
Brett House (17:06)
Yeah, and you're talking about that you're talking about the technical supply chain not the media supply chain like the programmatic real-time real-time bidding supply chain you're talking about the technical the tax
Crystal Wallace (17:14)
Well, I mean, the more tech stack that support that, but I mean, we transact on CPMs and you know, so what if, you know, and typically the FTE model is like a percent of, it's usually a commission based fee. So what if we've gotten to, you know, something that was more of a token fee for outputs and performance or something like that. It could be tech, but it could be some way in how we transfer
Brett House (17:40)
Yeah.
Crystal Wallace (17:43)
or we transform the current FTE model to something that is tangible, outcome-based, and not based on how many people you have on the account, but the usage.
Brett House (17:56)
Yeah, it's usage based. you and a lot of the AI companies and I'm kind of starting my own AI powered company. A lot of the talk is around that. It's around usage based pricing and the outdated nature of kind of seat based SaaS pricing, which to your point, you if you see that's the one big economic like sort of metric that you look at with with SaaS is that when usage goes down, no surprise, right? Churn goes up, right? There's a direct causation between and and you mark you track thresholds, right?
Crystal Wallace (18:11)
Mm-hmm.
Brett House (18:26)
And so we know, and I've known this throughout most of the SaaS companies that I've worked for, that once usage gets below a certain threshold, you can reliably and confidently predict that that company is likely going to call you at some point and say, we're not going to be renewing this contract. And so you're always, yeah.
Rio (18:43)
We're not renewing, yeah. Or we're renewing at a very steep discount, right?
You're not going to get a renewal plus whatever, and it's not happening.
Brett House (18:51)
Yeah, so you said platform-based world, which I thought was interesting because ⁓ maybe this leads us to that conversation around ⁓ gentrification versus platformization. and I just had a conversation about the CDP world ⁓ and the magic quadrant that came out that sort of spifurcated the two worlds and showed ⁓ two really different paradigms, which I think you were pointing to. Is it a platform-based world or are we moving away from?
Crystal Wallace (18:51)
Yeah. Mm-hmm.
Brett House (19:20)
platform-based world would be my question.
Crystal Wallace (19:23)
you always have to have an application layer, right? What does that application layer look like? Now, ⁓ I think for right now, it's still very much is ⁓ because...Brands want to control their data. The most valuable data asset now is what enterprise brands stated because the internet has been scraped by these LLMs. So they should be guarding and securing that. You can do data shares, can do cloud to cloud, you can do MCT to MCP, but we are still humans. We still need to touch, see, feel, light, sound, motion, what we're doing, what we're activating on across those different use cases.
I, let's not call it a platform, but you're still gonna need an interface. What that interface evolves to, not sure. The LLMs are making it such almost a wall garden, right? You can do, now we got ads, it can go and search the internet for it. They're trying to keep you in that environment, but.
It's like for me, like sometimes I want to be in more of a, you know, a chat GBT type interface versus, but sometimes I want to go into Excel because it feels nothing to, there's something, there's a behavioral science to like, again, format. Will that change? Sure. As you know, further adoption happens, people get smarter and you know, those habits, you know, kind of transform, but yeah, I think there still will always be an application layer, at least for right now. I think the internet of things is probably a real possibility. And so maybe like our traditional SaaS based platforms and things like that, that's not what we're talking about in terms of interface. It's the wearables, it's the augmented reality. it'll be, ⁓ now I'm going into futurist territory, but. ⁓
Brett House (21:18)
Yeah. Yeah, yeah,
Rio (21:28)
I like that. Keep going.
Brett House (21:28)
I love it. You're the technology futurist, but you're right, there is still an interface. That's what you're saying. Like whether the interface, whatever the form factor.
Rio (21:38)
But counterpoint, how much of that is just we've been trained to expect there to be an interface versus there actually is a need for one? I don't think it's going to completely go away, but is it minimized in the future?
Crystal Wallace (21:55)
I'm not sure. I am a format based person. So I am biased towards the way my brain thinks. And I think there are a lot of people who it may go in the future, but are we looking at five years? Are we looking like 20? Everybody got used to the smartphone really quickly. And so the forcing function of that was just ubiquity.
I'm not sure about this. Not everybody has the same access to. My LLMs are a lot better than most. Not only because I have the enterprise ones, but guess what? I actually spend about $150 and could spend more if I did more and weren't so stressed then. But just on premium subscriptions, on every, I have every major. There's all like, lovable, replic.
Rio (22:31)
This is true.
Crystal Wallace (22:55)
And I don't even use it because I don't have time, but I just want to be first to market. I'm a first adopter. so again, once it becomes critical math, and I'm not a scientist, I'm not that smart to predict those behavioral element things. But I don't think it's within the nexus, call it two years. I think we're still going to need platforms to interface with.
Rio (23:24)
I don't know, my hypothesis is that the future, like the eventual design, the UX of AI is this. This conversation. think, it's Star Trek. It's computer, make me transparent aluminum, like in the...
Crystal Wallace (23:35)
Yeah.
Brett House (23:35)
is conversational. Yeah, it's Star Trek, basically. You just talk to the computer and it does a bunch of...Yeah.
Rio (23:45)
in Star Trek 4, right? That's what I think it's going to be like. mean, there might be a ways off. ⁓ This is true. But I guess going back to like the discussion where I kind of pulling us back in a little bit. So, Chris, the one thing I'm really interested in, right? Like you talk about platformization to create of these like whether it agentic lay workflows on top of it. think for agencies, this is really interesting time. So I'm seeing all over the agency world, both Indies and Holdcos making really big investments in audience platforms and workflow management, brief creators, kind of building out this agentic layer. lot of that's happening at our Holdcos as well. It's like really exciting stuff that almost seems to be a different way of going to market. Let's go to market with IP, with platforms, let's be a product company. Love your thoughts on that, what you're seeing.
Crystal Wallace (24:13)
Yeah.
I think it's the right move. have to be, you know, have data and technical maturity in order to reach our brands consumers at the pace and play of just this highly, you know, changing world that we live in. ⁓ And first, you know, giving that power to your people so they can do things smarter, faster, you know, be more strategic for the brands. I think it's awesome.
It helps to bring all of the different workflows, pun intended, across Omnichannel. So creative, media, paid, search, own, marketing, know, all these different, ⁓ and again, I tend to go back to, where does the revenue door come in? You know, typically you have a CMO, you have a CFO, know, and they work across multiple parts of the, I would say, the marketing foundation. They typically have different influencers, key buying factors. When you have a common way to bring the function together, i.e. through these agentic workflows that are interconnected, that drive messaging bus within the environment, but also within the clients, it creates a flywheel of just common understanding it decreases a lot of the friction and tension that sometimes naturally exists between different, you know, cross-functional groups that are, you know, essentially all kind of charged with getting closer, driving more ⁓ equity and outcomes for the brand. So I think it's great. I've again been around the block and I've seen kind of
Brett House (25:58)
Yeah.
Crystal Wallace (26:25)
you know, the pitch theater of these platforms. And I could say what we have here at Omni, like it actually works. People are actually using it. It's actually solving the pain points. I remember when I ran a data science team, when I used to be a media buyer and had to go into like 10 different platforms and, you or you have to do all this rule-based logic. ⁓ To your point, Rio, we have a...an orchestrator at MCP like Swurm and I can just type from, give me some competitive intelligence. I can go on my specific module and sometimes I do just cause again, I like to tool and I don't know, be, I don't know, get my hands dirty, but I can do all functions from that one prompt window and it's pretty amazing. And so I'm just, I'm,
Brett House (27:19)
Yeah.
Rio (27:22)
It's better than 10 platforms, right?
Crystal Wallace (27:25)
10 platforms, know, having a brief, you know, 10 partners, you know, I mean, so yeah, it's, I'm excited.
Brett House (27:25)
Well, it-That you, I think you've been, it's funny, and I love how you think, Crystal, because you have me creating visual flywheels in my brain, right? You're visualizing, you must have been really good at math, because this isn't the secret to math, to visualize, almost in three dimensions, even four dimensions, and that's how you actually solve complex math problems, is through visualization. know, think Stephen Hawkins and his theories on the black hole, it was all diagrams and visualizations, but ⁓ to your point about,
Crystal Wallace (27:54)
discrete apathetic.
Brett House (28:01)
the functional, I think what you're talking about is GTM compression, right? It's customer lifecycle compression. So as opposed to having multiple functions that are focused on different steps in a process to delivering, let's say a product launch program.
Crystal Wallace (28:06)
Mm-hmm.
Brett House (28:18)
that's to, you for PNG that's going to market and there's multiple steps from sort of audience and ICP to creative development to media buying and segmentation and all that stuff. If you bring all that stuff together, all the different functions and people that are working separately under the FTE model suddenly can see it all within one model, which becomes a flywheel. And they're able to see a holistic picture of really it brings the customer in a weird way. It brings the customer, it sort of centralizes everything into one one play, right? Yeah, and it's, you would think that the tech might have the opposite effect of sort of pulling you further away or disassociating you further with like the human, but in reality you're seeing, you're kind of seeing all of this stuff and its impacts across a life cycle of different actions that the agency's taken to deliver a campaign on behalf of a product, right? Is that the right way to think about it?
Crystal Wallace (28:49)
You swarm around that cuff area? Absolutely.
Yeah. And I love how you said the life. No, I think that is absolutely the right way to think about it. And I love that you said life cycle because now we're not just campaign planning. You are planning for the life cycle of that customer value and even looking future proof to inform R &D type of decisions. What are the byproducts that a brand maybe can create another product on? Because now you're operating kind of in this continuum. So yeah. I think you know that at least the way I view it.
Brett House (29:46)
Is that what you guys are thinking about in terms of, because I think about when I look at the agency model and I think at the sort of the FTE based hold co model, I think a lot of the young independents that are very tech forward and Justin Townes Copeland and I, the CEO of 4A has talked about this at Adweek, ⁓ how they are as reliant on FTEs because they've become so tech forward, right? Do you think that that raises ⁓ a threat to the larger? ⁓agency holdcos in the sense that you can have smaller, nimbler, faster moving independents that can really move up the value chain and win very big accounts. Do you think that's going to have a competitive? ⁓
Crystal Wallace (30:30)
until they can't, until they can't. So I came from a startup, Search Ignite, Ignition One, that then got acquired by Densu. And we did some amazing things. We were absolutely tech forward. I won't report all the amazing things that we did, but there comes a point that you need clout. Now, back then, we didn't have kind of the acceleration of change in these, know, the agentic force that we do now, but...I think that there's, they're a threat, but are they a sustainable threat if holding companies get it right in terms of the right optimal mix of FTEs, people at the right altitude and technology enabling that. Does that answer your question?
Rio (31:20)
Yeah, it's interesting. my, my hypothesis, I was, that was good answer. I think my hypothesis, I agree with you, Crystal, is that I think the agentic or AI disruption for like ad tech and media planning and buying is going to happen really at a kind of at a platform level. So who's in platform today?
Crystal Wallace (31:34)
Mm-hmm.
Rio (31:36)
I mean, brands, unless they have big in-house teams, they're not doing it. It's really happening at the agency Holdco level. Who's managing most of the media spend? It's the Holdcos, right? So think that it's their opportunity now to adopt these tools and build it out for cost savings. And clients are going to expect downstream savings and better performance and all that. They're going to expect that as an outcome. But I think that's where the innovation happens because I mean, it's, I think MarTech's different than AdTech too. Like MarTech, sure, you can like build a gigantic workflows on top of a pattern. But AdTech really is, if you want to impact like,
So like a lot of it really is internet scale, right? So I think, so, so I think that like, it's going to be hard for these platforms to roll out things. So it's going to happen probably to hold co-level to people actually in platform every day, building these things across platforms. So I think that's probably what's going to happen. So I think that, I think it actually does ha being, being, being in a position where you manage a lot of spend and you have a lot of clients and you're actually is probably advantageousness. I would guess.
Brett House (32:26)
Yeah, it isn't that happy. sorry go ahead crystal
Crystal Wallace (32:29)
No, I was going to say, mean, your ability to view much more ⁓ observations in terms of what's really happened across the space. Yeah.
Brett House (32:41)
Yeah. And across their media spin and across their data, right? And the more, I think that's where you saw, you know, the creative agencies split from the media agencies, you know, 15 years ago or 20 years ago, which I think is you're to see that coming back together in a very strong way. And I see a lot of independent creative agencies that are struggling because their whole sales model was a linear model project base. We deliver a project done, but guess what? We're not, we're not serving the media. We're not collecting the data or orchestrating the data, which is what
Crystal Wallace (33:03)
Mm-hmm.
Brett House (33:12)
you guys what you've been doing your whole career crystal right and the more of that you do the more capable you are of competing on an IP level like you've got the client IP at the golden goose all in one place
Rio (33:24)
or Brett, even training the large language model. So you think about it, like I heard an interview like this, Dwarkesh Patel, like he interviewed Elon Musk this last week and he asked him, he how did you train your AI? He said, Tesla, we had all these people with little computers driving around the country, taking in all this information. So I think that the scale is a huge advantage for really building intelligence.
Brett House (33:37)
Yep.
Crystal Wallace (33:47)
Yeah, and you need scale to feed these models, so absolutely. I had a plan, I completely forgot it, but it was around...
Brett House (33:47)
Yep. Yep.
Crystal Wallace (33:59)
Just again, being able to also understand the ecosystem, not all these players are the same. ⁓ Not all the clouds are. You have to connect in a different way. When do you do a ⁓ Delta share versus, you know, and what my hope is is that how we're transforming at Omni, we can pivot whether we need to kind go against the big guys. We know what that competition looks like, but we're nimble enough and dynamic enough to compete with the independents just as well. And I'm seeing that sometimes feel sorry and just want to acquire and like, you know, give them the amplitude that they need. But I think what I'm saying is if you're playing it smart, you can actually, you can compete up and down, especially when you have to have a for your weight.
Rio (34:52)
Well, look at Omni, mean, maybe you return to that in Axiom, because I mean, you know, I was so excited about this Verger. The biggest reason was Axiom, I'll be honest. It's just, you know, I just thought of it the crown jewel of IPG and like one of the most valuable like data assets in the industry, right? So I'd love if you could maybe, I don't know how much you're able to share. I mean, obviously you probably have some trade secrets you don't want to tell people on a podcast, but how much can you share about your roadmap, what you're doing with AI, any agenda kind of...
Crystal Wallace (35:13)
Yeah.
Rio (35:19)
like layers you're building on top of that, how that relates to Axiom and your data set.
Crystal Wallace (35:24)
Yeah, so I can share a little bit. I'm not sanctioning here, kind of more as an independent, but it's super exciting. Axiom is the oldest people-based identity really in the world. We are global. Our real ID powers up, you know, the Fortune 500. We have nine out of the 10, you know, top banks.
So we understand how to move in a highly regulatory or regulated industry. Same with pharma and healthcare. ⁓ And just the ability, know, again, we know how reliable cookies are and even device signals. When you understand kind of at the personal level, and of course we, you know, we anomalize that when we need to across, you know, the various different channels, but it creates a precision, ⁓ a persistency and ability to really ⁓ drive efficiency and effectiveness and ultimately outcomes and ⁓ Omni and the tools and the data that are plussing up from, know, like Flywheel, the commerce data, like it's just going to be incredible. ⁓ Best identity spine, best commerce data, best, you know at the to know you know ⁓
Brett House (36:49)
Okay, we got a rule on signal and noise, no shills. No, I... Yeah. ⁓
Rio (36:53)
Hahaha!
Crystal Wallace (36:54)
I'm just saying, this one is worth it, but
Brett House (36:57)
Yeah, and I know the language you're speaking, having come from New Star and TransUnion. We actually had this conversation a bunch at CES, was this notion of identity and quality data being critical, especially persistent. And oftentimes it is still identifier oriented when you think persistence or semi persistent, because it's the stuff, identifiers that are less likely to change over time that you can still link back to a person in their household and that sort of thing.
Crystal Wallace (37:26)
That's right.
Brett House (37:27)
⁓ What do you think about the role of synthetic data and the ability to, you talked about this a bit earlier about, it's something that I'm doing with my company, it's sort of competitive simulations, buyer simulations, meaning like in my particular case it's B2B and you're going at a CMO ⁓ buyer. ⁓You know, how do you simulate ⁓ based on their use cases, their needs, their psychographics, their, the market intel, how they're going to respond to a specific message, a specific creative, a specific, you know, value prop.
Crystal Wallace (38:02)
⁓ So this is my favorite, like my favorite topic right now is synthetic audiences and building and playing with them. think they'll be more of ⁓ that like in coming or what have you. ⁓ I believe in the scientific method, so that's testing, testing. And so it gives us a way to test hypothesis and observability. ⁓ But again, with anything else, what...those synthetic audiences are trained on, it's going to really predict how valuable they are, how reliable they are. Again, a brand's first party data and those interactions, being able to put the right data in. And it brings together the quant fall, primary market research with addressable. So it allows you to do that stuff. ⁓
Brett House (38:41)
Yeah.
Crystal Wallace (38:58)
without having to go out and have a whole focus group, we can do it faster. Not to say that that goes away, but again, it's about the data inputs, but it's super exciting. I think it raises understanding quicker about who the true audience is. It definitely helps with innovation and helping you to reach those consumers in a way that drives kind of the equity of the brand, understanding kind of when and when you shouldn't be addressing them and know, work is going mass. So I'm excited. I even created a sona song about digital twins. So I'll share it.
Brett House (39:42)
And it's. wow, well yeah, and this, you think about it from your perspective, from an operations and tech perspective, I mean think the key here, at least going forward, is that you're not doing this on a one-off basis, like a campaign, it's much more systems thinking. You're not going, we're gonna do this for one campaign to ensure that the likelihood that this campaign engages and converts is X versus Y, right? That's great, but that's like sort of in a box, that's sort of time bound. I think the, yep.
Crystal Wallace (40:10)
It needs to be continuous and always on. And then your campaign are, you know, they are, they are a segmentation of those data inputs. And so you can say, this campaign really responded well. Here are the tactics, here are the partners, here are the activations that maybe you should do again to continue to reach. that's something that you need to be able to.
Brett House (40:29)
Yeah, the segments, the sub segments, Yep.
Rio (40:31)
Yeah,
I bother. I'd really like the example you gave of the synthetic focus groups. I actually saw that in action recently. It was really impressive. It was a, it was built for a client where, I mean, I've had clients who've done market research forever, right? A lot of my pharma clients way back in a day, he used to always do these, he would take these studies would take months and months and it'd be very expensive, right? And they would get a little book in the end, right? And then I saw in a matter of minutes, the client said, these are the different types of people I want in the focus group, different types of personalities you could pick. You could build your focus group and it would rip through this in seconds. was really impressive, time-saving. I wanted it little skeptical because I've always been a little skeptical of market research to be honest, being someone who comes from more... I I come from...
Brett House (41:19)
⁓ that's a whole other topic, Crystal. We got to tackle that one. ⁓
Rio (41:23)
another topic, right? Because I've seen a lot of dubious studies, I'll be honest, right? But this is super impressive. I was like, wow, this saves time. This is much, you know, it's less expensive, it saves time, and I thought the output was outstanding.
Crystal Wallace (41:27)
Fair enough.
Brett House (41:37)
Yeah, depending on the inputs for how you're testing, like if you're using some scientific ⁓ approach to personality, like the Jungian personalities, you can subdivide that in a thousand different ways. But qualitative data and survey-based data, I've talked to lot of data scientists, data analytics professionals, even AI professionals that continue to this day as much skepticism, Rio, like to your point that I've had about people filling surveys out, you ⁓ think about statistics and how
Crystal Wallace (41:48)
Yeah.
Brett House (42:07)
people sort of game like the Yelp and the review type of platforms because because you know you're you're likely to respond to the the the the people that love that love the experience or hated the experience but nothing in the mean nothing in the middle how do you think about it is and I don't know if surveys kind of follow some of that statistical logic but I think you can kind of run a regression and eliminate the outliers
Crystal Wallace (42:10)
Mm-hmm. Yeah. Yeah, mean, no, survey is a good one. mean, I think you have to design a survey right and you have to design it with built-in bias. And sometimes there does have to you.
They're typically going to be incentivized, but then you can back up and qualify that data with different mechanisms, different, you know, truth sets to ensure that, you know, there's not too much bias that's introduced. I think that they're extremely valuable. It can't be the only thing. It cannot be, you know, but I think that coupled with other, you know, ⁓ advanced analytics and, you know, data science ⁓ helps to... validate those panels. But you need to, you have to design it right.
Brett House (43:18)
Yeah, so qualitative data still plays a role, you know, it's agentified as what you saw, Rio.
Crystal Wallace (43:24)
Yeah. Well, yeah.
Rio (43:26)
Yeah, well, this is interesting. We've had this debate a couple of times, this pod crystal. I'd to maybe get your thought on this too. I don't know if this is right or wrong. That's right. But the debate was, okay, as we switch to this more like AI first, agentic world, you know, as we talked about before, maybe the whole UI will change or I don't think it'll be a prompt, right? But it's going to disappear. And like, guess the challenge we had, we have a great interview with a UX, like a UX leader who designs AI systems.
But it's like, does anyone, no one knows what it's gonna look like. does even, like, I look, I've worked in UX and CX for 20 years, right? A lot of the things like ethnographic research, ride alongs, different types of studies. Couldn't you even apply it to this? Or is it gonna be more like you just need someone to say, have a vision, I'm gonna do it, I'm gonna put it out there? I don't know, like, is a lot of these, are they still relevant? I mean, what's your thought?
Crystal Wallace (44:14)
you that, I think there's arguments that some methods aren't relevant because consumers don't act that way anymore. So 100%. That is a good provocation, not to say any of your others for it. Or maybe it has to be, well, no, yeah, we need to look at it. Actually, I'm gonna go and see which ones, if the normal kind of know, consumer journey changes. What does that mean for kind of, you know, that qualitative research? I think it just, I mean, does it go to just brain sensors? Like there's that, there's that type.
Brett House (45:00)
Hahaha
Rio (45:00)
Yeah, yeah. don't know. look,⁓ my hypothesis for, it's funny. I always used to get asked when I'd pitch clients on CX, this is like 10, 15 years ago, they'd why should we do this? Technology always changes. And my response would always be technology changes, but the needs don't. The needs will stay static and it's extrapolating what are those needs? How do you translate those needs into things, right? Into features, into products is, and it's something needs to change. So that was always the answer. But I don't know. Like I remember when AI first came out, it was like,
I almost had to step back for a few months and say, crap, like this, this might actually change everything. And I'm not, I don't know if I'm that convinced anymore, like about that we can't do it, but it's, but I don't know. I mean, it really made me rethink those, those, those methodologies that I learned over 20 years.
Crystal Wallace (45:45)
Yeah, and maybe it's the methodology is optimized, what really changes is the way you collect that signal, that input into those studies.
Brett House (45:53)
Yeah, yeah, yeah.
Rio (45:55)
Hmm.
Brett House (45:58)
Yeah, so I got a topic that I know you love, Crystal, is governance. data governance, it's so critical. I you talked about the middle layer with MCP. There's semantic layers built into ⁓ some of these platforms that help tell the AI basically what to do and how to respond. And it gives sort of guardrails, of decisioning guardrails. So governance, and that leads to data governance. It leads to some of the privacy conversations with trust, right?
Brett House (46:27)
Is that a, are there technical answers to this question? Is it policy based? What's your perspective on that?
Crystal Wallace (46:33)
technical policy based unknowns. That's why I don't think we can get rid of the human in the loop as fast as people think because we don't even know right now what governance challenges. mean, yeah, we have like, know, it becomes actually, you know, there are plenty of examples that are kind of scary right now. Imagine that at scale when everyone is using it. And so.
Brett House (47:02)
Give us an example. Give us an example that's scary. Do have one?
Crystal Wallace (47:04)
There... Well, I mean, when brought posted on X, users who... I mean, that's one. They have one where... Then there's some that, like, they've let the AI, like, especially now in research mode, you go and let it go off your research, and if you...
Brett House (47:15)
Yes, that was a big one.
Rio (47:17)
That was bad, yeah.
Crystal Wallace (47:27)
if you can give access to your systems to actually do things. And so it sent these emails that were completely inappropriate with clients. you you have things like that that have happened. ⁓ Data leakage, like it needs context. It needs, you know, ⁓ validations, not just of accuracy, but also from a governance and policy standpoint. And there are players out there that...you know, alongside with Axiom that are, you know, creating the frameworks, just like we had, you know, data governance and privacy, you know, it needs to be kind of the same like approach. ⁓ There's always been concern that you don't want to be too heavy handed or you stifle innovation, but this is one where, ⁓ you know, erring on caution, I think is really important. ⁓ We've worked, and this is why we are more of an ecosystem partner.
I think modularity and things of that nature are really important. I love MCP frameworks because I want you to do it all your agents and APIs and, and, know, validate that they are compliant governed, et cetera. I'll do the same with mine. We'll have ways to test those out, but again, you know, that's, those are, those are operating models and you know, partner frameworks that I think help protect with governance and then there's companies like IBM that actually they put together some really good AI governance around this and are playing more in that space versus, ⁓ although they have Watson too, which is great, but that's the way that, I think that there's a lot more and everyone needs to double down on what is this name? What is your compliance by design, your governance, specifically around these AIs. And it is, challenging because you want to release out to your workforce, out to your team so we can do things faster and better, but there's a real risk. those risks are, one of those risks are more costly than a whole bunch of little risks, so.
Brett House (49:50)
Yeah, it's like if you're too heavy handed, you sort of stifle innovation, you slow things down. But if you're not centralized and there's no enforceability, you're sort of expecting people just to adopt these best practices and these frameworks, right? ⁓ Out of the goodness of their heart, and there will be bad actors. So you sort of, you run into that risk, right?
Rio (50:12)
Yeah. It's interesting like being on the, let's say the selling side of a lot of AI a lot of times that clients ⁓ almost won't fund anything unless there's an AI written into it. But once you add AI into it, you get all these different groups who suddenly need to have a say, right? It's like suddenly governance is there, legal compliance, right? Which they should be.
Right. So, but I think it's interesting how I like that's what we need for funding now, but it slows things down dramatically and it's much more complex of a sale because you have all these all these nuances. Right. And regulatory landscape is changing.
Crystal Wallace (50:39)
Thank you.
Yeah, it is. It is. It is changing. I mean, there's rules that aren't even written. you know, whereas GDPR, CCPA was cumbersome. mean, we're going to deal with it and rightfully so. It's just we, you know, if you are a, you know, if you're a technology company or product company, you want to be AI for, you got to get ahead of it and help set and inform.
Policy and that's not to your benefit. It's what is right for you know, the world really in the consumers and so You're a good human
Rio (51:24)
But it's interesting, I think that kind of takes us back, Crystal, to your point before about like brands first party data, right? Because you think like you said about the LLMs, they've scraped everything, right? Anything that's public on the internet, they've scraped it. And it's unfortunate for publishers, they're not going to be compensated for that, right? But as I remember Brett, when we talked to ad tech therapist, Scott Messer, he said anything like that, that shit, yeah.
Crystal Wallace (51:40)
Yeah.
Brett House (51:48)
One of the many therapists that we've talked to in the Attic Martech space.
Rio (51:50)
That, yeah, had have a martic therapist and an addict therapist. His point was the ship sale. That's content scrape. Anything new, however, that you own that, you should monetize that. You should be paid for that. But the point about brands, like that is their data. It hasn't been scraped. It's yours. You should protect it. It has value. You can do things with it. got, obviously you want to do it in a privacy safe way that respects people's rights and their consent, but that there's a lot of value there.
Crystal Wallace (52:05)
Mm-hmm.
Rio (52:18)
in terms of building out their future, whatever the future is going to be, that should be the foundation of it.
Crystal Wallace (52:23)
Absolutely. And it does start with a data foundation, all of this. And if you lock in those AI governance guardrails at your data foundation level, then I think you reduce your risk. It's a risk mitigating function. And it's exciting. I actually like all the info set questions that I've seen in the new business pitches. There's more know, see-so, be-so type of security and audits and, you know, STEERS and data privacy assessments and ⁓ being a formal, you know, former, having an undergrad in risk management and insurance and, you know, an actual, it feels good to kind of see that, although it does slow down the deal, but it's the right thing.
Brett House (53:15)
So looking ahead, so I'm going to do one more futurist question for you, Crystal, because I think about this all the time and I don't think any of us have an answer for this, but it does seem like, and I think the topic around governance, that tech is moving at a pace that human institutions are struggling to adapt to ⁓ and can't keep pace with. And we talked about it in our last ⁓ podcast where the introduction of the printing press to Europe via China ⁓ led to the Reformation, right? And then we talked about the impact social media on our own children and how that sort of ⁓ it's sort of out piecing human
Rio (53:50)
Yeah, and in radio, like helping give rise to fascism, which is so, yeah, we talk about those trends a little bit. Yeah.
Brett House (53:54)
Yep. Yeah, and it's sort of technology outpacing human change, right? And even though it has a long-term benefit, arguably, ⁓ in the short term, it can have this really disruptive feel to it. I mean, it seems like we're moving very fast. mean, the product life cycles, the release life cycles, the innovation life cycles are so compressed, it seems, with AI. And we all grew up through the internet revolution, right? We're Xers, right? So we created the internet, right? ⁓
Crystal Wallace (54:18)
Thank you.
Brett House (54:26)
Now you're seeing the pace of change speed to a point where I think it's having some crazy effects around the world. My biggest concern is, ⁓ not biggest concern, but I always say this and I think it's theoretical, but it's not more than like a decade away, is what happens when quantum computing ⁓ becomes mainstream.
Right, right, when you take instead of having to kind of terraform your landmass to support server farms where you can do a trillion times the computations on them. Yeah, it's going to go into space on them and be powered by the sun, right? And you can you could just the computation ability. You know, we're talking about this race for compute power, right? With all the big ⁓ players like like with with with
Rio (54:59)
I was gonna go into space, Brett.
Crystal Wallace (55:02)
That's new one, right? Data warehouse.
Brett House (55:17)
This type of computing, the compute power is outraged. Didn't Google solve a problem that they would say ordinarily would take? I think they said, and I'm not being hyperbolic here, like a trillion years.
I don't even know how you, know, to solve because of the length and complexity of this. And it's a math problem that goes back to like very early days and like the beginnings of math that, and the Google test quantum computer solved this thing in two weeks. And I'm like, imagine, I mean, do you think that that has a, yeah, so that's the futurist in me. Like, how do we adapt to this change? And we know it's going to be within our lifetime.
Crystal Wallace (55:57)
yeah. ⁓The way you adapt is that you first you have to you know it Technical literacy needs to be just a common thing I think it's really important that one thing that we have lost with you know all of the advancements thus far especially with the younger generationals, you know them being able to contextualize Them you know the common sense and so some of those things need to be brought back so that you can run the quantum computer and the quantum computer can work for you. ⁓ I do think that we get to a point and you kind of just, mean, or the computers will control us. ⁓ But I do think that we get to a point where your ideas can be executed and essentially it can run your life.
Brett House (56:35)
Yeah.
Crystal Wallace (56:55)
these ages can run your life and you're just telling the ages what to do and you get to do whatever it is you want to do. think the counterpoint of that, like, do we go into a society that just really is our gluttonous and don't do anything other than like, you know, I don't know, feed themselves, like it's just, you know, cause we're not going to be.
Brett House (57:17)
Welcome to the United States. Come on.
Crystal Wallace (57:21)
Well, can you imagine that getting worse? Like where we're just like a yeah, so I don't know. I don't know.
Brett House (57:22)
Aren't we yet?
A sloth, a slothful human race. But so do you think this, so do you think a lot of these advances change? Cause we've talked about, and I think this ties back to a theme of like people need to be, you know, from our college grads, our kids that are, got a guy, guy at Syracuse and one that's about to go to college being, you know, the minute they enter the workforce, being much more technically astute and educated, because that's going to be a critical sort of knowledge base that you have to have to be able to harness the power of some of these capabilities, right? Does that change?
Crystal Wallace (57:57)
Yeah, for sure.
Brett House (57:58)
How does that change the rule of like the CMO and for brands that you guys support? How do you think about that, right?
Crystal Wallace (58:07)
I think the CMO role changes to that of a It's still about reach, it's no longer, again, it's not about campaigns. Maybe they become more of a behavioral scientist, right? You gotta really get in those mindsets in order to reach a consumer in this modern, agentic distraction of an age. They have to find where those new places are, and that takes them being much more of a technologist and much more scientific around you know, neuroscience, right? And yeah, at least having the constructs and the understanding of those very ⁓ technical or scientific things, they're gonna have to know enough to be dangerous and dictate the machine or direct the machine.
Rio (59:05)
I think the CMOs end up being more connectors and orchestrators, right? I I think that you're not going to need to be in platform to do a lot of things. I think not the CMOs were, but I think that also like, AI is going to be doing a lot of things we do today. And then I think being able to connect across the enterprise connect is going to be more more important. mean, CMOs, think if they stay in that silo and they expect all this work with their agency, they'll do everything and they're not able to work across the, working across the organization, enterprise organization effectively. It's going to be tough for them.
Brett House (59:34)
Yeah.
Crystal Wallace (59:34)
And I think those roles, like you know how during cookie deprecation and just like the acceleration of platforms, you have the new cheap data role, cheap growth. And so there was these bifurcation of roles at the C level. I actually think they get consolidated. So your CMO.
Brett House (59:50)
Yeah, those rules seem inherently-inherently CMO oriented, maybe without the in there, right? Customer and data orientation, right? And I think that power to superpower, the superpower I think is your ability to scenario plan to your point around a leveraging science to be able to test and learn at a compounded rapid rate that they just haven't, and that's why you need to be technical. You need to know how to harness the power of this agentic future to be able to test these things at scale.
Crystal Wallace (59:57)
Mm-hmm. Mm-hmm.
Brett House (1:00:23)
actually have the ⁓ repeatability in the IP, so it's constantly learning, which is what you guys are doing for consumer brands. I think CMOs need to do that within their own brands so that they're building this constant learning system, and it's refining itself, because that's going to make all of their decisions much more predictable for better outcomes. ⁓ Yeah, so they do act, to your point, Ria, as an orchestrator, a conductor of these capabilities that they don't you know because ordinarily it's just it's a little in it's a little market research there's some gut gut feelings and then
Rio (1:01:01)
And they're going to have to hire really good people who understand like no one can know everything. It's just impossible now. Right. So think they're going to have to like, need to have a really, if I don't know MarTech really well, I better hire a really good marketing technologist to my team if you're a CTO. If you don't know analytics well, better. I mean, I'm not saying they haven't done these previously, but I think it becomes more important to have those really like people who are deep in these areas.
Brett House (1:01:06)
Yep.
Crystal Wallace (1:01:17)
for you That's right. That's right. And the advice I would tell my grown children now and my granddaughter is go get you some some of those hard skills, know, the plumbing and and I mean, we'll have robots to do that. But right. But that that clip is we have more time on that than, you know, some of these other, you know, space roles.
Brett House (1:01:21)
Yeah. Yeah, that's, I love it! Hard skills.
Rio (1:01:47)
Yeah, well, Crystal, think that's a good point. people like, yeah, AI is going to do a lot of things, right? But like you better be really good at like reading, writing, understanding what it's what it's creating for you so you can make sure that it's like it's it's actually quality instead of just pasting something that's like if this then that and bullets. I mean, if people can see right through that, obviously the output of LLMS is getting better and better. But I think being able to discern what's good.
Crystal Wallace (1:02:16)
Yeah.
Rio (1:02:16)
edit things like you so it's almost like the humanities are making a rebound right there's core skills about understanding thinking philosophy of the philosophies rationalize things and being able to read and write well I have a stick those skills are coming back
Crystal Wallace (1:02:32)
They are. My biggest fear was all of this. your degrees are going to finally pay off.
Brett House (1:02:33)
Really, it's good for us. That's like our backgrounds, right? That
Rio (1:02:34)
Yes!
Brett House (1:02:40)
comparative literature degree, man, is going really far.
Rio (1:02:43)
Ha
Crystal Wallace (1:02:44)
My biggest fear is that, you know, something happens. We're not like a COVID and we're not able to access, you know, the models anymore. And we go into, we're so overly reliant on technology and there's no, there are single points of failure and we go into the dark age. There, always say there's nothing new under the sun. If you look at history and like the evolution of intelligence, both from like when you had scribes and only the church knew how to write, you know? And so, I always say, go back to history. And so we complain and be ready for this amazing celebration.
Brett House (1:03:16)
Yeah. Yeah.
Rio (1:03:25)
Well, well you think about this so the
Brett House (1:03:25)
Yeah.
Rio (1:03:26)
epic poems like the Odyssey of the Iliad Beowulf used to be memorized and and spoken by a person for eight hours ten hours over several days, right? I'm so that no one can do that now or think about it now like I can't even remember a phone number, right?
Brett House (1:03:32)
Yep, hand it down.
Crystal Wallace (1:03:42)
No?
Brett House (1:03:42)
Yeah, Herodotus, wasn't he the historian that brought a lot of that great knowledge from the Far East? Yeah, he brought it to the Western world, which had gone dark. The Western world had gone dark. We needed...
Rio (1:03:48)
of the oral mythology, right? He brought it together. like, but once... ⁓
Brett House (1:03:58)
knowledge from the Arab world, the Muslim world, that they had mastered math in algebra. And literally Europe was in the dark ages and we needed a translation layer. And Herodotus sort of brought a lot out history, a lot of those books.
Rio (1:04:10)
But, but, but Brett, but Brett, you think about it like, ⁓ pre-writing, people could memorize a 20 hour song, right? Once writing, once you began to write that, that skill went away. Like growing up, I still remember the telephone number of my best friend growing up, right? But like, I can't remember if you're number now in a million year, I can't remember people's names now, right? But the thing is like,
Brett House (1:04:18)
Yep.
Crystal Wallace (1:04:28)
Thank
Rio (1:04:33)
Like now that we have phones, that skill has disappeared. So I mean, like, and I read this study recently about people's brains over the last 2000 years slowly shrinking because like we've outsourced a lot of these things as part of our culture. So you wonder as AI does more things for us and it becomes less required to think hard about things. I don't know. I mean, like, I think that does impact us.
Crystal Wallace (1:04:54)
does and we've seen it I mean even just with these digital natives like these younger generations and I'll probably get dinged but the
Brett House (1:05:05)
There's no do you anything goes on signal a noise You will not get cancelled crystal
Rio (1:05:10)
Yeah. ⁓
Crystal Wallace (1:05:13)
Again, like, EIQ is like, I don't want to do this, don't cancel me. Back in the day, I'm always feeling like my grandparents now. It's just, it's like almost everything that comes out my mouth, feel like, you know, well, back when I was younger, but at the end of the day, I do think that there are effects that, you know, the cognitive thinking, that the emotional intelligence, like some of those things we have to work at them. I hope the educators out there thinking about curriculums to maintain those critical functions that make us human.
Rio (1:05:56)
Yeah, but it's funny you mentioned that. remember I had a, uh, someone on my team, just, will not reveal the anything about this person that could be attached to this person. Remember someone on my team who had an MA and an MBA, no, an MS and an MBA, highly and had worked agencies for several years who couldn't write. Like literally could not write a complete sentence. I could not put something that could be showed to a client without being completely edited. And I remember I said to this person, like, this is like crazy. Like, has anyone ever told you, you can't write?
Crystal Wallace (1:06:10)
Mm-hmm. Mm-mm.
Rio (1:06:26)
And the reply I was, no, like I've always been told I'm a great writer. So I said, well, you can't write. Here's elements of style. You put it in your back pocket and you're going to carry it around for the next six months. This person has much learned to be like a decent writer and actually thank me later. But it is crazy how, I mean, I, maybe I sound like that grumpy old man about the damn kids today. Yeah. But, but I think skills like that. I would argue now that like with AI, they're getting more important. Like we need to focus on them more. Cause I think I agree with you that a lot of that is maybe.
Crystal Wallace (1:06:29)
Ha ha!
Rio (1:06:54)
been lost to certain extent and we need to refine it.
Brett House (1:06:57)
Yeah, and I think those are hard skills. Like people say, you know, the reading, the writing, you know, those are soft skills. You know, computer programming, that's a hard skill, but you know, AIs proved us wrong there. That hard skill got completely disintermediated. Computer programmers can't find jobs, right? But the reading and writing actually is... Yeah.
Crystal Wallace (1:07:07)
Yeah. Yup.
Rio (1:07:14)
The good ones can, but I say like, you know, the entry level ones is harder, yeah.
Brett House (1:07:17)
Those are hard skills that the critical thinking is a hard skill, right? The systems thinking, the writing skills, those are all actually hard skills that are fundamental to surviving and thriving in this era. So I think that's a pretty good conversation. I we've been well over an hour. I think we're at hour 15. And thank you, Crystal. Yeah, it's getting late on a Friday and we're all running on fumes, I'm sure, because we all said it's been
Crystal Wallace (1:07:38)
people screaming at me shit. ⁓
Rio (1:07:45)
Yeah, this is all of us standing between us in a happy hour,
Brett House (1:07:48)
Yeah, yeah, my brain is ⁓ can't take anymore Patriots baby, I'm a New England Patriots fan and Rio is a Jets fan. So if this will be Yeah, he's like Yeah, the Sam Dardle experiment right so ⁓ Yeah
Crystal Wallace (1:07:50)
Super Bowl weekend, who do you guys got?
Rio (1:07:51)
I know. I Anyone but the Patriots. The Patriots losing is their Super Bowl.
Yeah, go Sam Darnold. I mean, we wasted, I can't, that's another podcast, Brevin, about like the Jets ineptness.
Brett House (1:08:12)
Yeah, well, and I got a text from my teenage boy saying, and it was the whole controversy about why Drake May did not get the MVP because somebody, an Irishman from Dublin, ⁓ voted for Justin Herbert as the MVP for, and it was like, what? So that was the one vote that actually changed. It would have been a tie.
Crystal Wallace (1:08:28)
Ha!
Brett House (1:08:35)
between ⁓ Stafford and May. So anyways, I'm a big Patriots fan looking forward, I actually think it's gonna be a closer Super Bowl than people think.
Crystal Wallace (1:08:36)
Thank
Rio (1:08:43)
Well, Brett, at least you are from New England, right? Because in my opinion, the worst type of human is someone who is a Patriots fan, but not from New England. That is the lowest form of life.
Brett House (1:08:50)
Well, it's like growing up in...
Crystal Wallace (1:08:56)
You see, I hadn't said who I'm going for.
Brett House (1:09:00)
Who are you going? Are you going for New England? Are you a Falcons fan? You're from Atlanta, right?
Crystal Wallace (1:09:05)
I'm from Texas, but I live in Atlanta. I'm a Texan, yeah.
Brett House (1:09:07)
you're from Texas. So, are you a Texans fan?
Crystal Wallace (1:09:12)
I would say I'm a cow, you could say I'm a ride along. I'm that despicable human that Rio talks about. Whoever's the best of the best I'm going for. I have to say probably cowboys when they're doing well. Obviously Falcons I'm cool with. I actually, when I was little, I was a little oilers cheerleader. So maybe type.
Rio (1:09:15)
Cowboys, okay, that's fair, yeah.
Brett House (1:09:16)
Yeah. Yeah. Yeah.
Rio (1:09:32)
All right, nice.
Brett House (1:09:33)
there we go. Yeah, I jumped on, I think football leans itself to bandwagonism, we'll end with that. I jumped on the Steelers bandwagon, because I was like, it was the 70s, right? And Terry Bradshaw and Franco Harris and you know, you're just like, that's what you saw.
Crystal Wallace (1:09:41)
Yeah.
Rio (1:09:47)
The Immaculate Reception, right? Remember that, yeah.
Brett House (1:09:49)
the Immaculate Reception, won four Super Bowls in my earliest years. so I became, I had this association with black and steel curtain, Minjo Green. I had this black and gold, I started liking teams that were black and gold. was this weird color association that I was like, this is hardwired into my brain. And then I had kids, which are now adults, Crystal. And I'm like, I can't, I can't, I've never even been to Pittsburgh. I'm like.
Rio (1:09:56)
Steel curtain. Hahaha You've never been to Pittsburgh. ⁓
Brett House (1:10:17)
I've never been to Pittsburgh, so I'm like, gotta raise some Patriots fans. I just, that's where the rest of the family is. So anyways.
Crystal Wallace (1:10:22)
There you
Rio (1:10:23)
Well, I don't know why I picked the Jets. I could have picked the Giants. I moved to New York and I was seven and like, it's like a bane of my existence. They're terrible. They're so awful. It's like the most poorly run organization, 15 year playoff drought. Like we're a farm team for other teams, like Super Bowl winning quarterbacks. So I don't know, it's bad, but that's another podcast.
Brett House (1:10:32)
Yeah.
Crystal Wallace (1:10:33)
I was just there. Yeah. Yeah.
Brett House (1:10:45)
Well.
Crystal Wallace (1:10:45)
Wow, I'm your commitment.
Brett House (1:10:46)
This has been awesome and thank you for the free flowing conversation. It's kind of like a jazz band over here. We set a little structure and then we let that we let the players improvise. So thank you Crystal for improvising with us and offering your insight intelligence and and it's been great. And for everybody that is listening and has made it all the way through this episode subscribe. Again we're on Apple podcasts we're on YouTube we're on Spotify wherever you listen or watch your podcast.
Crystal Wallace (1:10:57)
I it. I love it.
Rio (1:11:15)
smash that subscribe button, sign up, follow us, whatever it is.
Brett House (1:11:19)
and we'll see you soon. Thanks everybody.
Rio (1:11:22)
So long.
Comments