Dave Penski | Publicis Groupe | Media in the Age of AI | Part 1

Summary

CMO Confidential

Episode: Dave Penski | Publicis Groupe | Media in the Age of AI | Part 1
Host: Mike Linton
Guest: Dave Penski, CEO of Publicis Media
Date: March 24, 2026

---

Episode Overview

This episode of CMO Confidential dives deep into the rapidly shifting world of media and marketing agencies with Dave Penski at the helm of Publicis Media. Host Mike Linton and Dave discuss how AI is driving fragmentation, changing client-agency dynamics, challenging the creative process, and upending traditional agency economics. The conversation highlights how Publicis is responding strategically and technologically to these powerful changes, and what it means to lead in the world's most scrutinized marketing chair.

---

Key Discussion Points & Insights

1. The Business & Media Marketplace: Volatility, Fragmentation, and Disruption

Timestamps: 03:00–06:25

• Overall Environment:
  • Dave underscores ongoing tough conditions: volatility in the stock market, wavering consumer and advertiser confidence, and uncertainty around economic and political forces like tariffs and elections.
  • “We have a lot of volatility and just how we have lack of consumer confidence, we have a lack of advertising confidence and business confidence in what’s happening.” (Dave, 03:24)
• Marketplace Fragmentation:
  • Traditional digital duopoly (Google & Meta) is losing share as spend diversifies to platforms like Snap, TikTok, and LLMs, making media buying more complex.
  • “It used to be… you’re not going to get fired for buying TV and search… now we’re seeing that even those guys are getting disrupted by a much more fragmented marketplace…” (Dave, 04:38)
• Challenge for Agencies and Brands:
  • Increased layers—aggregators, delivery services, retail media, direct-to-consumer models—insert distance between brands and end-customers, making clear targeting and ROI measurement more difficult.

---

2. Agency Landscape & Publicis Strategy

Timestamps: 06:29–10:00

• Agency Market:
  • Publicis and Omnicom now dominate, with each having over a third of the U.S. market; strategies differ, especially around investment.
• Publicis’s Strategic Focus:
  • Heavy investment ($12B over 10 years) in data, identity solutions, commerce (e.g., Sapient, Epsilon, Mars United, Influential, Captivate).
  • “Our point of view… is the view of connected media. We believe in the end… that the connected identity is at the center of this and that we need to have an identity solution that allows clients to understand who their prospects, who are their current customers, how to target them and how to target them across paid media, CRM, influencer and commerce.” (Dave, 07:38)
• Contrast with Competitors:
  • While Publicis reinvests in tech and talent, competitors have leaned into stock buybacks.
  • “Our competitors have taken a point of view of doing stock buybacks which is not what we were trying to do.” (Dave, 09:23)

---

3. The Centrality of “Connected Identity”

Timestamps: 09:29–11:51

• Definition & Implementation:
  • Epsilon’s identity spine: 255 million unique U.S. IDs with associated physical and digital contact points.
  • Enables unified targeting and measurement across media types (from Hulu ads to influencer campaigns, CRM matching).
  • “If a client has a CRM file, we can match with them on a one-to-one basis… we then can track to see if they actually made that purchase.” (Dave, 10:41)
• Market Implications:
  • Growing realization that knowing who you reach (and with what message across channels) is now central to competitive advantage.

---

4. Agency Economics & Market Valuations

Timestamps: 14:14–19:10

• Performance Divergence:
  • Publicis reports solid growth and robust employee bonuses; competitors struggle with declining performance and stock volatility.
  • “We ended up just below 6%. We’ve forecasted growth between 4 and 5% this year… this is like year four or five of this kind of growth.” (Dave, 14:29)
• Valuation Challenges:
  • Despite strong results, agency holding company sector is undervalued, partly due to being “grouped with AI losers,” large amounts of company debt, and financial engineering (stock buybacks not creating true value).
  • “We are the nicest house on a bad block… we’ve been grouped as a category with some AI losers… but we actually see a lot of positive on the AI side.” (Dave, 16:42)

---

5. Insourcing, “Housings,” and the Shifting Client-Agency Relationship

Timestamps: 19:29–26:42

• The Insourcing Trend:
  • Many clients have experimented with bringing marketing and media functions in-house, but results are mixed; some move back to agencies (“out housing/rehousing”).
  • “We have 300 enterprise clients, we have 300 different setups… We’ve lost almost zero clients in this process… It’s a bit of a give and take.” (Dave, 20:59)
• Key Learnings:
  • Most clients end up with hybrid models.
  • Outsourcing remains attractive when in-house models grow too complex, expensive, or hard to scale and retain talent.
  • “It is a rounding error for us on a yearly basis. The amount of revenue we lose from in housing and don’t make up from rehousing ends up being a pretty flat number.” (Dave, 22:17)

---

6. AI’s Impact on Agencies: From Creative to Operations

Timestamps: 26:42–32:53

• Unprecedented Experimentation:
  • “We have thousands of AI pilots going at any one time… I have a lot of positivity in terms of this being able to transform our business in the next few years... The roadway will sometimes take longer… 95% of pilots failed in the first two years.” (Dave, 26:49)
• AI in Creative & Production:
  • The traditional “Don Draper model” is obsolete; asset creation (especially static and now video) is faster, cheaper, and far higher volume, although risks fragmenting brand coherence.
  • “The ability to produce 10,000 assets for the cost it used to cost us to do 50 is… convoluted to the actual end customer…” (Dave, 28:49)
• AI in Planning & Engineering:
  • Budget-cut planning that once took weeks now happens in hours, with full scenario analysis.
  • Major upgrades in technology infrastructure are now possible in record time—using AI for code refactoring, system upgrades at a fraction of historical cost.
  • “The ability to do those infrastructure and capex projects at not only a third of the cost, but a third of the time…” (Dave, 30:33)
• Media-Buying & Measurement:
  • AI/automation in media planning & buying has existed for years, but “agentic AI” is unlocking advances in strategy, audience building, and performance modeling.
  • “The amount of AI that is consumer facing today, that’s unbelievable. The amount that’s enterprise grade is still in process.” (Dave, 32:53)
• Caveats:
  • Barriers remain: legal considerations, first-party data protection, client-by-client readiness, and “hygiene” in AI training data.

---

Notable Quotes & Memorable Moments

• On agency economics:

  “We are the nicest house on a bad block.”
  (Dave, 16:42)

• On the AI hype vs. reality:

  “Often we make these predictions... but then there’s a long time between that prediction and reality. On the insourcing I would say we’re in year 10 of this.”
  (Dave, 19:41)

• On the speed of AI transformation:

  “The Don Draper era is beyond over... that can take 30 minutes. There are incredible tools that we are building... The ability to produce those images and ideas at speed and the ability to produce tens of thousands of assets across that...”
  (Dave, 27:45)

• On hybrid client models:

  “We’ve not fought this. We decided, hey, we’re going to support clients in each way they do it… If you decide to change your mind, we have a way to help you change your mind too.”
  (Dave, 26:02)

• Host’s summary of the market:

  “You have uncertain companies, consumers that are on a little bit of a roller coaster ride and a volatile marketplace... the segmented market that is getting more and more segmented is making that even harder.”
  (Mike, 05:54)

---

Important Segment Timestamps

• State of the business & macro trends: 03:00–05:54
• Agency strategies & Publicis investments: 06:29–10:00
• Defining “connected identity”: 09:29–11:51
• Agency economics & market valuations: 14:14–19:10
• Insourcing & “housing” trends: 19:29–26:42
• AI’s business transformation in agencies: 26:42–32:53

---

Tone & Language

The conversation is candid, pragmatic, and occasionally wry. Dave balances optimism about AI’s potential with realism about its “hype curve” and implementation challenges, while Mike distills complex agency dynamics with clarity and a touch of humor.

---

To be continued:
This is Part 1. Next week’s episode will further explore AI’s impact—stay tuned for Dave Penski and Mike Linton’s ongoing, unvarnished look inside the CMO suite.

Transcript

[A]

The CMO Confidential Podcast is a proud member of the I Hear Everything Podcast Network. Looking to launch or scale your podcast, I Hear Everything delivers podcast production, growth and monetization solutions that transform your words into profit. Ready to give your brand a voice? Then visit iheareverything.com welcome to CMO Confidential,

[B]

the podcast that takes you inside the drama, decisions and choices that go with being the Head of marketing. Hosted by five time CMO Mike Linton.

[C]

When is the last time you researched something on a website? If you're like most people, AI did that work for you. And that raises a question. If AI is doing the work, what is your website really for? This behavioral shift means AI bots are becoming your most important new visitors. A challenge our sponsor Scrunch is taking head on. Scrunch is the customer experience platform that helps you understand how AI agents experience your site, when and why they show up, and what's blocking them from being retrieved, trusted or recommended. Scrunch shows you the content and citation gaps and technical blockers and helps you fix them so your brand shows up when consumers Start with AI because your most important site visitor might not be a human for our listeners, Scrunch is providing a free website audit that uncovers how AI sees your site and how you're showing up in AI versus the competition. Run your site through [e-mail] CMO welcome

[D]

marketers, advertisers and those who love them

[C]

to Chief Marketing Officer Confidential CMO Confidential

[D]

is a program that takes you inside

[C]

the drama, the decisions and the politics

[D]

that go with being the Head of

[C]

Marketing at any company in what is

[D]

one of the most scrutinized jobs in the executive suite.

[C]

I'm Mike Linton, the former Chief Marketing Officer of Best Buy, eBay, Farmers Insurance

[D]

and Ancestry.com here today with my guest, Dave Penske. Today's topic Media in the Age of AI. Now, Dave has been in the agency business his entire career, including positions as the CEO of Zenith Optimedia, Chief Operating Officer at Publicis Group, and as the CEO of Publicis Media, twice a job he currently holds. This is his second time on the show. And full disclosure, Dave and I have known each other for a few years and our past also crossed recently when he ran Zenith and I was working as the CMO of Farmers. He has in my mind one of the best seats to see the tectonic shifts going on in the agency business and the media markets. Welcome back, Dave.

[E]

Thank you, Mike. Appreciate it.

[D]

All right, Dave, let's start. Give us a quick overview of the business marketplace, not even just the media marketplace, but the overall business marketplace and what you are seeing today.

[E]

So it's a great question. I mean, it's a, it's definitely. I've looked at it much more by vertical. So how are QSRs are doing, how a pharma is doing, how about insurance? And you've worked at a couple of different verticals. So I don't think it's a simple answer on anymore. But I would say overall it's a tough marketplace. I think that we see it across. We see it some on the stock market, we see it some on the, the, the earnings that have been coming out. It is a challenge marketplace. I think that we have a lot of volatility and just how we have lack of consumer confidence, we have a lack of advertising confidence and business confidence in what's happening. And it's not a political answer. What's going to happen with tariffs? They were shrunken down and raised all in one day. I think certainly those dollars could end up going back to companies and going right back into advertising, which would be very good for the advertising marketplace. At the same time, I think as we go into this midterm election, there's a lot of different promises of positivity and there's a lot of promises of challenges. So I think that we're going to continue in this kind of what we've been on for the last year, which is a very challenged marketplace. I'm sure you're going to ask me some AI questions coming up. I think one of the things that we're seeing that's really come through over the last couple of years are the amount of companies coming between our clients and their end consumers. So whether that be on the, what used to be much more direct to consumer, whether that be retailers and retail media, whether that be the delivery services and the QSRs, whether that be the aggregators and the insurance business there. There's just quite a bit of disruption. And I think the media marketplace is becoming more and more fragmented. So, you know, it used to be, you know, people say, you know, it's 20 years ago we, you know, you'd be a brand manager at a large cpg, so you're not going to get fired for buying TV and search. And then it was, you know, you don't get fired for buying Google and Meta. And now we're seeing that not those, even those guys are getting disrupted by a much more fragmented marketplace where an advertiser might have spent 50% of their digital money Between Meta and Google, those numbers are going down. Now their total spend is going up. This mean Google and Meta spend is going down, but their percent of their digital dollars decreasing because there's just so many more options. Whether it's Snap, whether it's TikTok, whether it's a new test in, in the different LLMs, it's just a very fragmented market and at the same time you have a very challenged market of clients looking for business results that are outsized their competitive set.

[C]

So, so I have, if I parse

[D]

all your comments, what I hear you say you have uncertain companies.

[E]

Yes.

[D]

Consumers that are on a little bit of a roller coaster ride and a volatile marketplace where you can have things like the Sass apocalypse any minute from a fake, fake research report. So what you have is it's really tough to figure out what to do. And the segmented market that is getting more and more segmented is making that even harder.

[E]

Those are all very accurate. Thank you for saying it much cleaner than me, Mike. I do appreciate.

[D]

Oh, I got the chance of hearing you so I just condense it on the show. So, so the agencies in this talk about the agency landscape, it looks like a really tough market.

[E]

It's a good day to be publicist today, so I can't complain. So I think obviously it's a pretty seismic shift in the agency marketplace right now. We have ourselves an omnicom and I'll speak for a US perspective who have become far larger than the other competitors. Both are over a third of the market. And you know, it's, it's a. We're, we've taken two very different strategies. Our strategy has been about growing through more M and A and investing back into our people, into our company and into investing really taking our dollars to reinvest back in, reinvested a lot in

[C]

data and the back end before anybody did.

[D]

Right.

[E]

Yeah. So over the last 10 years we've invested about $12 billion. I'd have to look at. I know the, I know the Euros have to think about the translation today might be a little different. So over $12 billion into from Sapient to Epsilon to the influencer companies. And really what our point of view of what we've been, what we've been talking about is the view of connected media. We believe in the end is that the connected identity is at the center of this and that we need to have an identity solution that allows clients to understand who their prospects, who are their, their current customers, how to target them and how to target them across paid Media, CRM, Influencer and commerce. And what commerce means for certain clients is different whether that means on Amazon on that or does that mean through other commerce plays like it could be if I'm talking to a qsr, is it through Doordash and Uber Eats? So all of those ability to go across those platforms is really important. So we've made a number of investments. We've bought a large commerce agency, Mars United, we bought the two largest by revenue or cost in influential and captivate in the influencer space. And we're making those all kind of stick together with the Epsilon identity spine and then the other probably through line. And it's interesting some of the companies that you've worked at and certainly we look at one of the more AI proof things out there is sports. So we've also made a couple different sports acquisitions and we've been open that we are continuing to be open for business and looking for places to continue invest if it helps us and our clients do those kind of five things. So the four plus sports are really where we are looking to be investing in the future. And so our competitors have taken a point of view of doing stock buybacks which is not what we were trying to do. So it's a very.

[D]

Your IR people will be so proud of you for this latest answer.

[E]

Yes.

[D]

So fantastic. I want to make sure everyone, I

[C]

want to go back to the just

[D]

general agency landscape in a minute but when you say connected identity, I think we all know what it means but you should be really clear what exactly when you. Yeah, identity.

[E]

Yeah. So for us, we want to make sure for each of our clients and for us that we're able to look at. We'll talk about, let's talk about from a US perspective but it's global too. There's some different rules by market but this from a US perspective that we understand who every person is. So we've connected, we use, we use Epsilon as our core identity partner and we own them. So be clear that we have 255 million unique IDs where we have address, home address, email address, phone. We've, we've caused that resolution go down. We know by device, how many devices, if you have five or six devices in your household, how do we actually know it's that one person? How do we target them from a CRM standpoint? So we can do a match. So if a client has a CRM file we can match with them on a one to one basis. We can do this with both PII and non pii Matching, we were able to do that across paid media. So if someone runs an ad on Hulu and we can see if we then expose them through an influencer that we have tracked all the influencers we then can track to see if they actually made that purchase. And we can also look to see are they part of our current CRM file and they're a current customer or are they a new prospect? So we start looking at audiences, how do we go about doing that? So that is what we have been focusing on and that's a really important part of our differentiation between us and our competitors. And I think that Omnicom is starting down that path on trying to get an identity solution with Axiom. And WPP came out and said they didn't need identity. Now they've kind of said they do need identity. I'm not sure of the latest WPP positioning on this identity, but I think that most of our clients have certainly come to the conclusion of if I can reach these consumers through CRM that are my current customers, or I can reach them very efficiently through influencers, I can reserve more of my paid dollars to reach my prospects who I need to convert. And it might be more difficult to convert over. And that's been something of how do you then you know, all of our clients have a limited amount of spend. I don't care if it's a billion dollars. Everyone feels their budget should be bigger. So how do we then use those?

[D]

Except the cfo, that's never.

[E]

That is accurate.

[B]

Now that take a short break from our discussion for a word from our sponsor, Scrunch.

[F]

Over the weekend I asked AI to do dozens of things for me. Please help me make this dishwasher start working. Settle a debate between me and my partner. Is it safe to sleep at 80 degrees?

[E]

It's not.

[F]

Please help me find a robotics kit for a five year old operating at a middle school level. And even help me compare the Scrunch Agent Experience platform to with Adobe's LLM optimizer. In every one of these interactions, I never once visited a website. In this day and age, what's your website actually for? The truth is, almost everything on your website is now for AI consumption. The Scrunch Agent Experience platform helps manage, optimize and deliver content directly for AI to retrieve.

[B]

Now back to our conversation with Dave.

[D]

And I want to make sure I'm getting this right because one of the things I hear you saying is you guys are leading. Actually, I know it's not exactly media,

[C]

but if I say media is My

[D]

ability to reach the target with the most certainty.

[E]

Yes.

[D]

The entire company is kind of leading with that first. Is that a way to think about it?

[E]

Yes. We want to be able to. To know who we're reaching with and certainly when we get into it, is, I think, where we have some of our largest upside. We've had a great. We had a great year, and we've. We've posted this. You know, I can talk about, you know, previous results. You know, we actually had one of our best years ever on creative and production. And we've seen an ability to not only know who that person is, know where they live, what, what, what actually motivates them and putting the right content in front of them. And if you look at what we're able to do from an identity standpoint with, not just from a media perspective, but that actually then produce the correct asset to put in front of that, that customer is. Is very important. So. Absolutely.

[D]

So let's talk about everybody that's not you. Yes, publicists.

[E]

I try not to bash my competition, but I'll do my best, but go ahead.

[D]

All right, well.

[C]

Well, just in general, the rest of

[D]

the forces, we know all the forces on the agencies. What's your take on these forces on the agencies and then in general, all agency economics. Yeah.

[E]

So, you know, we just went through our earnings. Obviously, we've shown a very positive margin and growth of last year. We ended up just below 6%. We've forecasted growth between 4 and 5% this year. We are. We've been on this trajectory for quite a bit. This is like year four or five and five of this kind of growth of this or higher. And it's been a. Certainly, we've seen it. We've announced to our employees, you know, a very positive bonus to go along with that. And certainly I think most of our employees will be very happy with those results. And I think that's something that we feel very positive about, that we want to pay our people fairly. We want to make sure that we're rewarding those that have helped us deliver these, These incredible results. And I think some of our competitors, I think WP was yesterday. They're. They're showing that it's gonna. They were, you know, minus a bunch last year, will be minus a bunch again this year. Whether it's gonna be high single digits around high single digits, it's going to be a challenge. I think it remains to be seen on. On Omnicom exactly where they stand. You know, their fourth quarter is their first quarter together. It was a bit of a confusing announcement. And you know, one of the things that's going to be challenging is, you know, we report net in gross, they report only gross revenue. So there's a lot of pass through, there's a lot of deals there. I think it's hard to really show, but certainly we have a much higher growth rate and certainly a much higher number of where we believe will be.

[D]

And we just had Pete Imwala on, he's a former CEO of rpa, a

[C]

big independent agency in la.

[D]

And one of the things he talked about was other than publicist agency valuations have really taken a hit.

[E]

Oh, the actual valuations, yes.

[D]

And insourcing by a lot of clients. What's happening even when it might not be the best thing for the client. Can you talk about those trends in the agencies and then also the l insourcing move?

[E]

Well, the valuations are, you know, we always say this in a politically correct way for us, but I think we are a, the nicest house on a bad block. So I think that is a challenging situation. Evaluations, we of course believe we should be valued far higher. I'm sure, I'm sure all of our competitors probably feel the same way. We've been grouped as a category with some AI losers. I haven't really seen that. I'm sure we'll get into that in a minute. But the value, we actually see a lot of positive on the AI side. But having said that, I think that the valuations are the valuations, whether we think they're fair or not. It's, it's what the marketplace is showing. It's a little bit confusing, you know, because of the, the amount of debt that sits on. You look at the total market valuation of wp, it's not as low as the, you know, as you look at the market cap, I think it's £3 billion today. It's, it's low, but there's, there's still a lot of debt in that business and there's some other things going on there. So it's not exactly a fair comparison. But, you know, certainly we would have expected our, after pretty stellar results for a number of years, a much higher evaluation. Now Omnicom has gone up from a stock price because they're, they're pushing a huge amount of buybacks into the system. But that's not bringing your market cap up, it's just bringing up the stock price. So.

[D]

Right.

[E]

That's going to be interesting to see. You know, that's not something we believe is A long term strategy is you're going to keep cutting.

[D]

Yes. You're trying to buy back your stock forever.

[E]

You can't buy back your stock forever and you can't just keep cutting to growth. So I, I don't think cutting 20,000 jobs is going to bring you growth. And I also think that we'll get into the client question in a minute. But you know, a number of our employees now, they of course had synergies so there's, there should be some value to bring these business together. But on the other side, you know, a majority of our clients are an FTE model that you can't keep cutting because then you don't get paid. So it is a, there is a part of this that, you know, it's, we're going to see how clients react. So these massive cuts and if they're going to client those, those, those savings are be passed on to the clients versus being kept at the holding company level. And that remains to be seen. But I'm certain if I was a client and my holding company was cutting 20,000 jobs, people are coming off my business, I'd want the money back. So I think that's a reality that we're up against. So.

[D]

And you guys have been rolled into. And you mentioned that the nicest house on a bad block.

[C]

I think.

[D]

Yes, that, you know, whole verticals are thrown under the. This is an AI negative or this is an AI positive. The agency business that's been thrown into the AI negative can.

[E]

Yes, absolutely.

[D]

And then you throw in the company's insourcing thing on top of that. Yeah, tell us about.

[E]

Yeah, I'll answer

[D]

the AI effect.

[E]

I will answer the, the, the insourcing one first and then I will. And the insourcing thing has been going around for a long time, but there's been a lot of like we make these huge announcements and then like nothing happens. I remember I was actually speaking at a public board meeting not that long ago and someone asked me a similar question. I said, you know, in 2004 I was told that would be the last upfront. The last upfront was gonna be the 2004, 2005 up front. And I just saw a meeting on my calendar for the 26, 27 up front. So obviously that's 24 years off on the prediction of that was the last up front. So I think often we make these predictions and that we don't think sounds great, but then there's, there's a long time between that prediction and reality. And on the insourcing I would say we're in year 10 of this where there's a significant amount of clients that have brought stuff in. We also have what we call out housing, which sounds better than it is, which is bringing it back to us. So I don't know if we want to call it outhousing. Maybe we come up with a better name. You're the cmo, Mike. That may be out a different word.

[D]

Rehousing.

[E]

Anything but rehousing. Doesn't sound very good either Mike. But so we have a combination. And I would say if we have 300 enterprise clients, we have 300 different setups. There are clients that will be in housing small amounts. Their belief and maybe correct actually is in order to do their larger data models, they need to own the data contracts. So they bring the data contracts in. Some clients are bringing the hands on keyboards in. Some clients are bringing in almost all of the investment, others are bringing all the investment with us and keeping more of all the strategy going in house. And I think our point of view is we want, we've not fought this. We decided, hey, we're going to support clients in each way they do it. We've lost almost zero clients in this process in terms of the bringing the whole thing in house. And what we've noticed is it's a bit of a give and take where often clients want to bring us in house. They want to hands on keyboard and then they go, well, we don't want to do the bill pay and we don't want that. That was terrible. Or we, you know, we'd like you to do that part. And also, you know, we can't get people to stay as it is hard and, and you know, very often they might not have a great, you know, person comes over does a year ago, well, I want to be promoted like, well, we don't have any other jobs. This is, this is it, this is it. So now we gotta do so but,

[D]

but what I hear you're saying between the.

[E]

Yeah, so I, I'm not. It is a rounding error for us on a yearly basis. The amount of revenue we lose from in housing and don't make up from rehousing out housing coming back, I think it ends up being a pretty flat number. And often clients that want to bring things in house, we have other services that we might sell them on. Hey, we can train your hands on keyboards, we can do that. So it's not a, this is not a drain on our revenue that I think we probably would have thought it was ourselves by the way 10 years ago when oh my God, we're going to have to reduce all this and we really have not, have not seen that as a thing. And you know, if you look at our revenue, it plays out that this is not then a huge effect.

[D]

And beneath this I hear you also saying there's no roadmap for all the housings in housing, out housing, whatever housing you want to say. And so there's no playbook that actually has caught on. What there is is a bunch of effort.

[E]

There's a lot of effort and we get to AI a lot of pilots. And whether or not, you know, some large enough clients, there is some, there's certainly some value that can be had. If you're using the in housing strategy for a larger business strategy or you are truly a marketing company for companies that aren't marketing companies and they're bringing stuff in house who have no experience doing it, it can be very challenging. But if you're a marketing company who might have a large retail media business as part of your overall business and you're running an advertising business today, it makes sense to bring a number of these things in house because it's actually part of you now becoming or part of your core business versus a client that might be, you know, as you were forming the insurance business. I don't think insurance business is going to start a retail media business. And so bringing in something in house to build out a large in house buying and and I think what the biggest challenge becomes often the job is tougher than it that it looks on paper. And if consultant comes in and tells you you can save money then you get into the details of can you really save money? And also are you getting. I think one of the things that we've seen is do in year one, your talent's great. Year two, who's going to retrain them? What is their career path? What is you know and often you're dealing with very. We're used to dealing with a lot of under 30 talent. Now you have this large group of under 30 talent that has to figure out a way that they want to get promoted, they want to get more money, they want to figure out their career and how are you then going to keep them? Now if you have a larger market we have clients like Walmart who are not only great marketers and great retailers, they're also quite an advertising company. Same with Amazon, Target and I'm not picking on any of mine versus others. The Best Buy, Lowe's, Home Depot, they all have incredible marketing Retail and advertising businesses that they're running. And for those companies, a lot of it makes sense to do in house because it's part of their core business offering, or if you're an entertainment company or wherever that would be, it starts to make sense. So I think there are places where it makes sense. But even all of those companies, half of those are clients, half are not. And we've helped all of them in terms of that. And none of them have brought all of it in house. And that's something in a number of our clients that brought some stuff in house six years ago are now in the. Because it's funny you look at it. Well, we have all these, you know, you start looking at. We've now often happens as we've all done this and we've done it too, is you hire 25 people to the job and all of a sudden you've noticed in the last six years that 25 people's balloon to 80. Like, how do we have 80 people doing this? Well, let's bring it out. And then you go, we could, we're paying these current people 20 million. What will you do? You'll do it for 10, we'll do it for 10, we'LL move it back out. And then all of those conversations are happening. I think that's something that is going to continue to be part of the ecosystem. But again, this is not something that we want to solve these problems for clients and work with them. And I think the one thing we decided 10 years ago is we were not going to fight it. We were going to do the opposite. Say this is what you want to do, we'll support you. Here's a bunch of our tools that you can still use. Here's a bunch of our things that you can still work with us on. And here's ways that you can, if you decide to change your mind, we have a way to help you change your mind too. And that's been a very positive outcome.

[C]

So let's, let's flip to the AI

[D]

part, the second part of the question and agencies and AI.

[E]

Yeah. So a couple things, you know, we have, we have thousands of AI pilots going at any one time and different agents that are being built. And I have a lot of positivity in terms of this being able to transform our business in the next few years. I do think that the roadway, just like the upfront, hasn't been dead for 25 years. The roadway will sometimes take longer, especially when you get into legal on some of these things. And I think there's an article from MIT that 95% of the pilots failed in the first two years. And there are some of these things that are going to take longer. There are, there's huge. I think we're, we're seeing probably the two pieces of massive disruption is on the creative and production business. The ability to create assets and you know, the old Don Draper is beyond over where you go send 15 people to work for two weeks on a project, they come back to the idea to go back and forth. That can take 30 minutes. There are incredible tools that we are building that we've partnered with. It's not my exact expertise. I won't say the wrong names here, but we have certainly what I've seen us been able to do. I think any of you who are listening, we. What you're able to do on a personal level on a ChatGPT or other language models, and you can imagine the ones that we're using for this are even better. And the ability to produce high quality static, especially video, not quite there, but getting there, the ability to produce those images and ideas at speed and the ability to produce tens of thousands of assets to be across that. The only thing that I would say is you produce 10,000 assets. The negative to the fact that you can produce 10,000 assets for the cost it used to cost us to do 50 is often it can become off as very convoluted to the actual end customer that they've now seen from one company, 14 different ads and different styles because they're all trying to hit in the right place. So there does take some discipline and work to do that. But so that's been one place.

[D]

Everybody is using AI.

[C]

AI becomes the average creative.

[D]

And so, you know, which works for maybe a lot of the creative, but maybe not the breakthrough creative. So this is super, a super interesting evolution.

[E]

Yeah, I think that very often as we've used many different tools over the last 20 years, they've all been additive. This one though, does feel different, especially in that place. The second part where we're seeing a massive advantage is the ability to do things much faster. So one of the things that you came to me and said, hey, you know what, we need to do a budget cut and it could be anywhere between 25 and $100 million. And that is okay. So while we come up with a 20, 40, 60 and 80, $100 million cut plan, I can give you fully detailed plan, all of those now in one day where that might have taken a team two weeks to do one of Those plans, right.

[D]

And a big team that's not like

[E]

two people and that could go. So the ability to do that more accurately and faster are dramatic. And then probably the place where we're seeing it, you know, we own Sapient. And I won't just, but I'll speak more to what Sapient's able to do. But it certainly is across on the engineering side where many of our clients have older code and, and things that need to be redone and you know, a lot of infrastructure projects that we're going to almost, it's overwhelming at four and five years to do them. The ability to do those infrastructure and capex projects at not only a third of the cost, but a third of the time because most of the challenge is just about the money. It's about, I can't have this network down if I'm running a rental car company for the better part of the next three months while you fix this thing. And I got, I got Mike coming

[C]

to rent a car as a rental car company.

[E]

Yeah, it's not good to be down. I mean we've all been to rental companies. It's quite, quite a great, great experience sometimes. But the, the, the, the ability to be able to rewrite code to fix this and to move at speed. And so I, I think from our Sapient business we see a lot of upside on the AI side. And I do think that there are going to be AI winners and losers in that piece. And the ability for us that, you know, we're not like some other managing consulting businesses, have massive, massive amounts of people. We are, do have a more technological view and have tech solutions for clients. So we're seeing some positive in this ability to actually do engineering at a much faster rate and to be able to do it at a much more efficient rate by using AI automation to rewrite and to update new code and you know, they think about the amount of our clients are spending to upkeep their current systems to be able to do that at a much more efficient rate. So I do think on both the engineering side and the creative side, we're seeing massive changes now. Media, we are seeing changes. The ability, kind of the ability to redo modeling, redo measurement, redo a plan faster on the other side, on a lot of the buy side, we've been using AI for a long time. It hasn't been agentic AI, but it hasn't needed to be and, but the ability to automate and optimize has been part of our lexicon and part of our strategy. For many many years. So I think you're seeing less changes on like the buy side but as you would on more of the strategic build the audiences those ability to kind of produce the different models doing MMMs at a much faster rate all those things but again those are the positive sides. The amount of time it takes to go from there to reality is differs by client. You still have to go through are we able what technology are clients comfortable with? Who from legal is signed off and bringing first party data into a large language model. All of those things are not simple questions.

[D]

Management and the people thing there's a

[E]

lot to it and I do think that it's going to take a bit more time on the other side. A lot of the geo stuff and a lot of the ability to fix some of the hygiene behind what the large language models are looking at and finding is stuff that our clients are doing. But yeah, I think that the amount of AI that is consumer facing today, that's unbelievable. Is there the amount that's enterprise grade is still in process.

[D]

Got it.

[B]

This ends part one of our discussion with Dave Penske. Join us next week for part two of our conversation.

[C]

When is the last time you researched something on a website? If you're like most people, AI did that work for you. And that raises a question. If AI is doing the work, what is your website really for? This behavioral shift means AI bots are becoming your most important new visitors. A challenge our sponsor Scrunch, is taking head on. Scrunch is the customer experience platform that helps you understand how AI agents experience your site, when and why they show up, and what's blocking them from being retrieved, trusted or recommended. Scrunch shows you the content and citation gaps and technical blockers and helps you fix them so your brand shows up when consumers start with AI because your most important site visitor might not be a human. For our listeners, Scrunch is providing a free website audit that uncovers how AI sees your site and how you're showing up in AI versus the competition. Run your site through [e-mail] CMO.