All About AI: Digital Agencies Reimagined in The Age of AI

In the inaugural episode of our “All About AI” series, CJ Bangah, Principal at PwC and leader of their US software and digital platform practice, joins Digital Clinic to explore the rapidly evolving landscape of marketing agencies in the age of AI and digital transformation. Drawing from nearly two decades of consulting experience, CJ discusses the changing role of agencies, the impact of emerging technologies on marketing practices, and the critical balance between AI-driven efficiency and human creativity in the industry. The conversation covers key topics such as the future of digital transformation, the challenges of data monetization and privacy regulations, and the shifting dynamics of customer journeys.

Throughout this first episode, CJ offers valuable insights on how agencies can remain relevant by focusing on strategic expertise, developing proprietary technology, and fostering a client-centric approach. She provides a forward-looking perspective on the skills needed for future marketing professionals and how agencies can thrive in an increasingly tech-driven ecosystem while maintaining their core value proposition. This episode sets the stage for our ongoing exploration of AI’s impact on digital marketing and business transformation.

CJ’s Professional Journey and Role at PwC

Aaron: CJ, tell me a little bit about your professional background. It’s very impressive. You clearly work with very interesting, very large clients in fascinating spaces. So tell me about what you do and how you came to be doing this work. 

CJ: I appreciate the question, and for better or worse, it was a happy accident. I come from a small, blue-collar family—teachers, plumbers, roofers, electricians. Finding consulting and having been a consultant for the last 19 years came through starting a job out of college. Somebody I worked with said, “What are you doing here? It’s a 38.75-hour work week. You’re here 50-60 hours, trying to make sure we ship new code on time, get everybody into the application. You really should be at the Big Four.” 

I had no idea what it was. I put myself through undergrad and couldn’t make some of the interview days that the big firms have. So it was literally just somebody that I worked with at my first job out of college that directed me to my former firm, and then I’ve been with PwC for 14 and a half years. 

Aaron: Your role with PwC is as a principal, and you lead their digital transformation practice. Is that correct? 

CJ: I have a few titles. We like to joke that it’s small hats, and you just try to keep them carefully balanced. I look after our software and digital platform sub-sectors. That’s big tech clients who have SaaS business models, ad-supported business models, things like that. I also help with the media sub-sector and author PwC’s perspectives on the future of advertising and marketing. But really, I’m just a non-CPA partner, so principal, who serves clients in tech and media, and I’m pretty fortunate to be in the space. 

Digital Transformation in Practice

Aaron: Can you give examples of the type of work that you do with clients? 

CJ: I’ll give you one of my favorite project examples. I had a client who faced declining revenue growth. They were launching new products into the market, it was getting more complicated for them to sell, and they had a profitability problem. They saw their competitors being much more aggressive in going after their client set, and they knew they had to break the paradigm between costs rising at a rate greater than revenue. But they didn’t want to do it irresponsibly, and they wanted to do it in a way that harmonized their global community. 

We had the opportunity to work with 60 countries around the world to figure out how they made things work, what was similar, what was different, and how we unified it. For a large digital advertising publisher, we helped them transform the model from nuts to bolts—new ways of working, better ways of collaborating, better ways of scaling. 

The reason I really loved that project is that I initially thought of it as rolling out a new tool and changing global processes from prospect to cash. But I didn’t understand the importance of engaging humans in a really deep and meaningful way from a change management perspective. I learned that when we’re doing projects, we need to do them with people, not to them, in a way where they understand why we’re changing, the benefit for them, and how new tools and processes can make their job easier. 

I also learned a ton traveling around the world. For example, I didn’t realize that some of the folks we talked to in Japan, if they didn’t fully deliver on their commitment, had to go personally apologize. So they would undervalue all of their contracts in terms of how much advertising inventory they were selling because they didn’t want to under-deliver. Once you understood that there was a deep cultural reason for it, it enabled you to make better design choices as you’re trying to get them onto global standards. 

Aaron: In that example engagement, were you able to work from strategic ideation all the way through fruition and implementation—strategy through execution? Is that typical? One of the knocks on consulting is that consultants come in and deliver ideas and then walk away. It sounds like that’s not the model for you at all. 

CJ: It sometimes is the model because that’s what clients ask for. But I think the best consultants are valued partners who care just as much about your business as you do, if not more, and they’re really committed to driving results. Ideally, they are there to make sure that you go all the way through to execution and you actually deliver on the promise of what you want. 

We even price some of our projects that way. It’s not a rates and hours conversation; it’s a conversation where it’s a value share, and we may get a bonus if we’re able to successfully deliver on the business case or the initial premise. 

Sometimes clients will say, “I’ve got great project managers, great developers, Six Sigma certified process re-engineers. I just want you to come in and inventory the pain points, pull in industry best practices, tell me what we should do, and then we’ll figure it out.” Sometimes they are successful. Sometimes we have to go back and say, “Hey, let’s help you get this back on track, clarify the questions, and make sure that everyone understands why they’re doing what they’re doing, the benefits in it for them, and the things that need to be true for it to be successful on final execution.” 

Aaron: Can we talk a little bit about digital transformation? It’s another big consultant buzzword that I get the sense isn’t very well understood. Can you unpack the definition of digital transformation and then talk about what it means in the real world in some of your engagements? 

CJ: It’s a good question, and it’s one that has been mercilessly abused by people who either want to weaponize it as if you don’t digitally transform, you’re irrelevant, or use it to drive cloud revenue or online revenue. I think the potential of what it could be often gets lost in conversation. 

Digital transformation is embracing the fact that since the internet came to exist, so much of our life as a consumer is grounded in these digitally enabled experiences, whether it’s a pure-play digital experience, a digitally augmented experience, or something where there are digital underpinnings to how you are connecting with the world. Those B2C experiences have a way of propagating into B2B journeys and experiences as well. 

If you’re using technology thoughtfully and prudently, and you’re getting rid of business models that don’t make a lot of sense, then you’re digitally transforming. You’re using technology in incredibly helpful ways. 

I think what we’re going to see with generative AI is a new wave of digital transformation. It’s not just about using great technology systems to underpin your business, having your systems connect, or upskilling your humans on how to use technology in highly productive and effective ways. It’s about having flexibility in your business model, where you’re able to light up and wind down cloud capabilities as you need it and as you grow. 

Gen AI is going to be interesting to see what it does to the lexicon and the taxonomy around the role that digital plays in our lives. The recent outage we saw through a deployed system shows that the dependence on digital technology does have risk and downside. We see that with social media, concerns about job loss with Gen AI, and a lot of different propagations of it. 

Aaron: What is your view of digital transformation? How do you think about it? 

CJ: My view tends to be very grounded in stronger tech foundations, a more productive workforce, and an elevated understanding of how technology can move the business forward with flexibility. But when you approach it, do you have a different definition? 

Aaron: That’s consistent with our understanding as well. Because we’re a performance-driven marketing agency, we tend to think about it very pragmatically. We’re thinking about the ways that we can employ digital systems, new digital technology, and new data to deliver business value as quickly and efficiently as possible. 

We tend, sometimes to our detriment, to come at it from a limited strategic perspective—strategic with regard to the channels and the campaigns that we’re managing. We sometimes forget to come at it from an organizational perspective and think about digital transformation for the organization overall, and the way that could, by extension, enable performance for the business. 

CJ: If we take my very broad enterprise-grade view and translate it to the marketing and advertising arena, I think that’s where things get really interesting. If you go back even five or ten years, one of the big arguments from an agency perspective or a marketing perspective was, “Are you spending enough online? Have you embraced the fact that consumers are spending so much time online, and digital underpins so much of their experience, that if your ad dollars are not effectively flowing to digital channels, you’re not actually effectively engaging your consumers?” 

Every year that passes, you continue to see digital advertising in the US and in established markets around the world grow at rates that are greater than other form factors. I think people are catching up. Out-of-home advertising is one of my favorites because if you have a digital out-of-home ad, it’s this incredibly powerful palette. Folks are not watching television shows at the same time consistently, right? So if you want to have a broad reach, digital out-of-home in certain neighborhoods or in many neighborhoods combined can actually be interesting to think about. 

The same goes for performance marketing if you’re looking at search and some of these other form factors. The digital transformation in advertising and marketing is interesting because it’s still hard to convince brands and folks that may not be technically savvy to fully understand the form factor. They may be worried about where their ad is going to show up and whether or not they’re really going to drive the ROI. Yet, the industry as a whole just continues to grow and have pretty impressive growth year in and year out, inclusive of form factors like video game ads and other things as well. 

The Impact of AI on Digital Marketing

Aaron: You mentioned several things that I’d love to dig into. First, AI is very much in the zeitgeist. When AI first came on everyone’s radar, the initial fear was that it would eat everyone’s lunch—it would be about cost reduction, elimination of jobs, increasing efficiency with technology at the expense of human beings. That has largely proven not to be the case, with some exceptions. It seems now, when I’ve heard you talk about this, that the emphasis and the value being delivered is much more about enabling additional revenue, enabling additional creativity. It’s an augmentation, as opposed to a technology that displaces humans. Can you unpack that a little bit? 

CJ: A lot of my thinking on this was shaped by the former CEO of IBM, who said, “AI is not going to replace your job. Other humans who use AI are going to replace your job.” I think for many industries, that will be at least partially true. There may be use cases where we decide that it doesn’t make sense for humans to do that work anymore, and an AI-powered robot may step in for certain parts of menial labor. There are going to be other jobs that are so fundamentally transformed that the volume of people needed could also potentially change. 

But the premise that AI will completely replace humans overlooks the fact that the amount of sophistication companies need over their data governance, data control, data harvesting, and insight creation to have really great generative AI is going to be a journey that takes months and years, not days or weeks. 

There are bespoke use cases and really fabulous creative tools where you can generate something AI-powered quickly and do a tremendous amount of experimentation. So there are really advanced Gen AI use cases—the conversational AI capabilities, the art rendering capabilities, the editing capabilities, and other things like this. 

But I think we went so quickly out of the gate with a lot of focus on how transformative Gen AI was going to be. Not enough folks were deep in it to see that it’s actually going to be a journey that looks like this: It’s not purely linear, and it’s not purely quick. It’s going to have a lot of different considerations that impact the pacing, including things like labor unions for the fields that have some of the more advanced, really clear Gen AI use cases and value propositions. 

Aaron: It seems to me that when we all started initially playing with Gen AI, it was very exciting that it could produce a result very quickly. Then it became clear that it could produce a result very quickly, but the results were usually bad. It’s usually 80 or 85% of what you actually needed, and the last 15% really mattered and took an exponentially greater amount of time, due diligence, investment, and refinement to get. That still seems to be the case. 

I’d like to get your perspective on this: My perspective is that if you’re dealing with a task, particularly in digital marketing, where the task has been or can be clearly documented and it is repeatable, then it is a task that is likely to be automated by AI. The value that remains in digital marketing, at least in an agency like ours, is in true strategy, deep expertise, creativity, and orchestration. It’s not really in implementation anymore. 

CJ: One of my favorite Gen AI stories came from a panel I facilitated at Cannes. It was about the opportunity with a creative asset, like, let’s say you’re advertising a car. You film the automobile going through a rugged terrain, whatever, but then be able to personalize that ad so that the car looks like it’s in snow or on a beach or somewhere that is going to more authentically connect with the viewer of the ad. That’s kind of neat. It’s kind of cool, and it’s not going to be appropriate for every brand. 

But I think that your point is, you know, first of all, for those that may not be as deep in the space, AI has been powering digital marketing for over a decade, and there’s a lot that’s been in place for a very long time now. Gen AI and the Gen AI use cases have continued to mature at an accelerated rate over the last year and a half plus. 

But I do think that understanding humans – being able to build a marketing strategy and a marketing campaign that’s deeply grounded in both consumer truth and brand truth, and matchmaking to find that right opportunity, and then using AI capabilities to optimize the tactics – is crucial. So you’re not having to manually look up a bunch of stuff on building audiences from scratch. You can have audience recommendation engines, you can have help with creative automation, things like that. 

With a human – I hate to say human insight engine, because it’s such a lame term – but with the right human engagement and understanding, that deep essence of what makes us as consumers want to connect with a brand and want to connect with a product or service, I think is really important to not lose sight of. 

Aaron: A few months ago, Sam Altman made a very provocative comment and claimed that in three years, AI would be doing 95% of what digital agencies are doing today. What’s your reaction to that comment? 

CJ: Whenever someone makes a comment that is visionary and transformative, you sometimes have to ask: Did they make that comment because they really believe it, or did they make that comment because it’s going to be something that grabs headlines and is provocative and gets people thinking? 

I have to imagine that there is a world where 95% of the work could be done by Gen AI, but if you were going to put it side by side—a 95% Gen AI-controlled digital marketing agency versus an agency really intentionally using Gen AI in the right areas—your customers aren’t going to want to call up a Gen AI robotic agent to explain to them, “Here’s what I’m trying to do. Here’s what my customers like. Here are the things that we’ve tried in the past and they haven’t been effective, or that we’ve tested and learned on together.” 

I think there are portions of the ecosystem that do not add value in the way that they should, and those portions could be automated and augmented by Gen AI capabilities. No question. But the things that do add value are going to be Gen AI-powered, not Gen AI-eliminated. 

Aaron: I think there are some agencies where 95% of the work they’re doing could be and should be automated by Gen AI. But that’s a different issue. Conceptually or categorically, what is AI good for, do you think, in the future, and what is it not good for? 

CJ: It’s interesting because sometimes I have AI capabilities integrated into my email, and it will draft the response for me. It makes me feel like I don’t care about the person on the other end if I use the Gen AI-enabled response. That’s my own change management learning curve—it’s actually just a productivity enabler, and it’s something that can and should be used heavily. 

There’s also the capability to transcribe your meeting, summarize your action notes—these are things that I don’t like doing and get no value from. By all means, embrace Gen AI use cases and value-add there. 

I also think if you’re doing quality control processes over financial statements, or if you’re trying to figure out the insight from the data, not having to manually read 10 or 20,000 rows of advertising performance information and having a Gen AI friend who makes recommendations to you so that you can be smarter and then ask more provocative questions and get to really great recommendations—there are use cases there as well. 

The business model reinvention portion of Gen AI is a little bit too early to say exactly how that’s going to play out, but I do think we’re starting to see companies think about data monetization. So I have this information on this subset of consumers—can I then get other people to pay me for this data on my subset of consumers so that they can plan their digital marketing campaigns, make them smarter, make them more action-oriented, where they have full-cycle investments? I think those are definitely on the horizon for the future, but it’s way too early to say exactly how it’s going to fully manifest. 

Data Monetization and Privacy Regulations

Aaron: That’s a perfect segue to the Internet Advertising Report, which you help to produce every year and is fascinating. There were three themes that were clear in that report, and the first has to do with monetization of data to enable much more sophisticated targeting. What do you see happening there? 

CJ: It’s interesting because part of me thinks we’re going to have a “data is the new gold” phase over the next few months or years, where more companies and industries are going to be really interested in how to use data in much more compelling ways to train their AI models, target their marketing efforts, and really get under the hood of this. 

There are also, according to the IAB’s research, about 2,000 privacy regulations globally around the world right now on who has ultimate power, authority, and control over consumer data, consumer data use, and how long a company gets to retain it. GDPR, CCPA, state-by-state privacy regulations, things all over the world that are coming in. 

Part of me thinks we’re at the beginning point of a data gold rush, where we’re going to see a lot of companies try to monetize their data, try to get access to other people’s data, reinvent their business models with it, and really try to get a lot smarter about how they’re thinking about investment of capital and investment in marketing spend. Then the regulators are going to come in, and it’s going to change and be transformed again. 

Another part of me thinks that the ability to fully and effectively monetize that data is going to be hampered by many of the systems that a lot of industries sit on. So it may not be a gold rush; it may be another hype cycle. I’m a little torn on the full-scale opportunity, but I do think it’s at least worth considering and exploring because the last thing anybody wants to do is waste money, especially money that’s designed to help you unlock growth and attract new consumers, which is the whole point of marketing. 

Aaron: As we’ve seen over the last couple of years, any significant shift in privacy regulations, particularly in healthcare with the broadened interpretation of HIPAA (which now is also in question because of recent legal decisions)… So everything is up in the air around HIPAA privacy regulations. As you said, there are 19 states that have their own privacy regulations. I think there are some 350 other potential privacy laws that are pending. 

What that suggests to me is that any forward-looking, any prudent business, should do everything they can to take control of their own destiny and not be beholden to third-party data, third-party cookies. They should develop a first-party data strategy, figure this stuff out on their own, develop the direct relationships they need to be effective. And when we’ve been able to do that, the performance is also much better as well. 

CJ: It’s interesting because if you go back to the 1960s, the content producers—and I’m using a media example—relied on someone else to be their distribution vehicle. So someone else knew which households the content was showing up in, how many customers were churning every month, how much they were willing to pay for a subscription, and potentially had more research on them beyond that. 

What happened with the internet and streaming is you started to see these direct-to-consumer businesses owning the entire value chain and realizing how incredibly powerful that can be. Not only are you reducing your dependency on somebody else for the success of your business, but you’re also able to get such a deeper understanding of your customers—who they are, what they care about, what they react to, what you’re not going to get any engagement on. 

Then you can get into the other part of a really great marketing motion, which is not only am I bringing new customers in the door, but when I get them, I delight them. I keep them. They’re loyal. They’re a referral source, and they are someone where I’m genuinely adding value to their life with my products or services, or at least serving some baseline need if you’re talking about some of the foundational companies out there. 

The Evolving Customer Journey and Attribution Modeling

Aaron: We’ve talked about two of the three trends that were highlighted in this report: one related to monetization of data, two related to increased privacy regulation. The third is this collapsing of the customer journey, the customer funnel, from discovery to purchase or consumption. Can you share a little bit more about your perspective there and what you’re seeing? 

CJ: It depends on if it’s B2B or B2C because I think it’s both collapsing and expanding depending on your customer set, your segment, what you’re looking at. Forrester came out with some really great research that showcased the average B2B customer journey was like 27 touchpoints. It’s everything from doing an internet search, watching a product video, talking to their friends, going to your website—all these different things that are like this new consideration pattern. And there aren’t a lot of B2B companies that can connect those purchase signals or drive the consideration. 

The same is true on the B2C side as well. You’re not necessarily going to know if somebody’s watching a short-form social video where they did a makeup tutorial and that drove them into the store to buy your lipstick, or if it was because you have a really beautiful product ad, or if it was because you did some awareness-building through digital video, or all of these other factors. 

I think owning the value chain, owning the consumer engagement, and then being able to adjust those signals back into your marketing and sales engine are incredibly powerful. The touchpoints in the journey can very much be compressed, especially if you look at shoppable ads, where the discovery to the purchase can happen in a microsecond. Whereas for other things, it gets much more expansive. 

But I think for digital commerce and retail media, that’s where you’re seeing the form factor grow so quickly. It’s measurable, it’s attributable, and it happens in the moment. You’re not waiting X weeks and collecting 20-plus data touchpoints to try to show the conversion. 

Aaron: It seems to me it’s increasingly problematic to even attempt to collect those touchpoints. We go back to privacy regulations and things like cookie consent and the fact that you can’t track most of those touchpoints now in many contexts. Which brings me to attribution modeling. Given this context, given the issues with privacy regulations and cookie consent and the multitude of touchpoints, how should we be thinking about attribution modeling, and by extension, modeling that enables us to understand and project where we should be spending media? 

CJ: Attribution modeling is a really interesting area, and it kind of depends on the brand, the marketer. You can get incredibly sophisticated—100-plus data points, fully fed model, throw in some great data science, some AI analysis and other things, and come up with a solution that looks pretty interesting and pretty beautiful. But will they trust it? 

I think the future of attribution modeling is something you probably have a view on as well. I’m imagining that what attribution modeling meant when the internet first came to bear and when internet advertising first started growing versus what it means now are also very different. The complexity that has come out of the last 20-plus years is creating an environment where attribution modeling is still expected and it’s still part of the ecosystem, but more companies that I work with or that I’ve seen are trying to simplify their attribution model while still collecting more rich data and insights, including their own CDPs and customer understanding and customer awareness. It feels like an interesting area that’s going to continue to have a lot of debate on the right way to do it. 

Aaron: We see the same thing. We have instances in which we’re using media mix modeling, and that’s very complex and sophisticated, involves data science, and even in the simplified, streamlined version, still takes weeks or months to build and get to fidelity. You have to continue to train and refine it, and that works, but the complexity also, as you suggest, can create skepticism. We’ve proven out the models, and that’s been good, but the level of effort is really significant. 

On the other hand, we have some clients where it’s just super simple attribution modeling—this signal, this result, that’s it—and that’s perfectly fine, and it turns out that that works very well. And we have a third case as well: because of HIPAA regulations, we lost all attribution modeling, or all attribution I should say, all tracking, and had to fly blind. We used our own models and forecasts, and they were as good as the full fidelity attribution modeling that we had before. 

CJ: I love that story. One of my favorite quotes from when we did the IAB outlook a few years ago was that the advertising industry can get drunk on its own data, and it’s never enough data. You just gave an example of when actually, sometimes you don’t need to bend over backwards and collect 500 data points to try to prove the value of marketing. You can come up with an opportunistic model or other different ways of looking at it. 

I also think occasionally, especially because the CMO-CFO relationship is one that could be filled with antagonism and lack of understanding, with the two executives speaking very different languages, attribution modeling is something that you could see being baked into the mind of a finance person and then used to hold the marketing person more or less hostage to showing the results and driving the value. 

One thing that we found with many of our clients is there are different, better ways to bridge the CMO-CFO gap and to look at it through new, innovative methods that may not be your traditional media mix modeling or may not be this overly complicated attribution model. What those are and what works for different brands and different industries can vary very much, but I do think that mindset of “we don’t need to fight with the finance organization to justify our spend or to prove out that every single dollar was perfectly invested” is one that actually can be transformative for the modern marketer in today’s age. 

Aaron: Surprisingly, in that case that I mentioned, it was the CFO who loved the modeling that we did the most, even better than when we could go in and re-implement analytics in a compliant manner. He was saying, “Why would we do this? We’ve got this model, the model works,” and it was the marketers who wanted the data. 

CJ: Marketers love data.  

Aaron: I completely understand. I am one.   

The Agency of the Future: Adapting to Change

Aaron: If we think forward, what sort of talent, what sort of people will populate a marketing organization, and how will that differ from the people who are there today? 

CJ: That is a fabulous question, and one that’s in some ways hard to answer because I think the functional boundaries that exist in marketing today—like the separation between your analytics and insights team, your AI team, your data analyzers, whatever their brand is, and your creatives and your media planners and so on—there are going to be portions of some of those roles that are augmented or enhanced with generative AI. 

Part of the question is, will it be in five years, and will it be enough that I don’t need two completely different skill sets for these two functions? I actually need people that are more well-rounded in their understanding of the entire marketing value chain and the entire value cycle. So the folks I’m recruiting in the future have a more cohesive blend of art and science, a more cohesive, robust understanding of the storytelling and the creative side and the data and the analytics side, which has been happening for years. But I think some of that could be accelerated slightly. 

When you’re looking at marketing talent five years from now, the marketing talent most likely is going to need to be much more experienced and aware of the customer and the customer expectations and the customer journey, and how to collaborate effectively with sales and finance and some of these other functions. They’ll need a more fulsome understanding of marketing as an art form and as a business growth enabler as well, which probably means the “I do this kind of marketing with this kind of skill set, and then I pitch it over the fence to this person who sits over here” model is going to start to become much less relevant. 

Aaron: I have wondered if the kind of person who will be needed, and maybe arguably is needed now, is someone who can function, can think beyond the constraints of current convention and process, and can think much more about what could be rather than “this is the way we’ve done it.” That seems to be the most needed skill, at least where I sit in our agency. It’s getting out of your own way and realizing, “Oh, there are 15 new ways to do this,” and you have to get really comfortable with being uncomfortable. 

CJ: I love that point. I also think the problem-solving in the future and the relationship-building to be good at it and to be able to get others to want to do what you want to do… The ability to use technology thoughtfully in appropriate ways to contribute to really great solutions, the ability to work across functional values—a lot of these things are going to become so foundational to bringing an innovator’s mindset to the work. 

Because, again, the pace of change from a tech enablement perspective has gotten to the point where it’s moving faster than we are. So I think the desire to learn and the desire to constantly be educating yourself and upskilling yourself is going to be core to every profession, but marketing for sure. 

Aaron: We need restless people. 

CJ: Exactly, those who are never satisfied with the status quo. Which, again, sounds kind of exhausting in some ways. And so then it’s the Gen AI use cases. I hope the most valuable are those that actually make it easier to be effective in the things that we do day in and day out and make our lives a little easier because I think some of the problems that we’re solving just get harder and harder. The number of people we need to solve them with continues to expand, and there are so many other factors that make the world more complex, so the more we can simplify, the better. 

What do you think the agency of the future is? 

Aaron: My hypothesis, and it has been for the last couple of years—and it wasn’t that I was prescient and saw AI coming, I just had some hunches—my perspective has been that mid-to-low-level digital marketing is rapidly becoming commoditized, and it’s not worth investing in it. The only thing that matters is strategic expertise, deep expertise. You have to change the agency model and not have five people with a lot of experience and 20 people who are junior, who do the work. You need 20 people with a lot of experience, who have elasticity of thought, who are creative, and technology is now table stakes for agencies as well. 

So a digital marketing agency that just does digital advertising and some digital strategy and maybe implements analytics—that also is commoditized. There’s a race to the bottom in terms of what people are willing to pay for that, and there’s a decreasing value in that as well. 

What we’ve tried to do is to develop our own proprietary technology that gives us a leg up, and to hire increasingly senior people, and to really focus on developing strategy as a muscle and an instinct throughout the entire organization. Then we’re using AI as much as we can. We use it in some of our practices; we don’t use it for direct-to-client deliverables, but we use it to augment deliverables. We use it for strategic exploration, or of course, things like synthesizing meetings and all that kind of thing. 

But we’ve also built custom GPTs that do some interesting things. Again, getting to my perspective that if something can be really rigorously documented, AI is good for that. We built a GPT that uses Google’s content quality rating guidelines—it’s a 160-page document that’s really detailed. It is a recipe, so we trained a GPT on it, and it does a really good job of assessing a website, as well as a junior strategist would. Then you still have to check it; you wouldn’t give that to a client, but it points you in an interesting direction and gets you 80% of the way there. 

So that’s sort of the perspective: It’s expertise and strategy, it’s proprietary technology that gives us a clear point of differentiation, and also, because we’re small, creates a justification for selecting us. So far, we have started to build our own creative team; that may or may not be the right path, given the evolution of some of the tools. But in digital advertising, in particular, creative seems to be the new optimization. We’re not going to optimize based on data in the same way. We’re not going to use automated bidding optimization. We’re going to use creative to optimize. 

Those are sort of the three pillars that I think are important for a digital marketing agency. And now I would love it if you tell me why I’m wrong. 

CJ: If AI is replacing the junior talent… 

Aaron: What creates the senior talent? I don’t know the answer to that. I’m not sure how you cultivate them if you aren’t bringing them along.  

CJ: For decades, advertising and marketing has been grounded on “I can help you plan and execute a beautiful campaign that’s going to drive your business outcomes because I know the consumer set.” But the world that we’re in now has such multi-generational entertainment, media content consumption differences—whether it’s the amount they do or do not play video games, the amount they are or are not spending in augmented and virtual reality environments, whether they watch short-form or long-form video. 

One of my favorite quotes from Cannes this year, just because it got one of my fellow colleagues really excited and then had a very disappointing punchline for him, was “Long-form video is back,” but what they meant by long-form video was six-minute videos. 

I think the modern marketer is dealing with a world that is significantly more fragmented and diverse in terms of the kinds of experiences that consumers are looking for. Some are based on age, but that’s even not a great proxy for it—the language it’s in, the culture it represents, the values it showcases. 

So if your junior talent is now being augmented or replaced by Gen AI, you don’t know where your senior talent is going to come from in the future, and you’re missing that “I’ve hired people who help me understand younger generational preference patterns.” How does the agency of the future close that gap as well? 

Mind you, that method has always been deeply flawed, right? Because we can’t look at the world that we live in and assume it’s the world that all of the target customers of every brand that you’re working with also are considering. There’s a lot more that should go into it, and historically hasn’t, so it could be a great forcing function. But what’s your perspective on that side of the equation? 

Balancing AI and Human Creativity in Marketing

Aaron: Well, I don’t have a great perspective on that, except to say that that aspect, the truly creative aspect of digital marketing, developing a new campaign, developing a new brand, is not something that we pretend that we do. Instead, when I say creative is the new optimization, I really mean we’re focused on creative to drive campaign optimization. 

When we work with clients, our reason for being, the reason that we have such long-term relationships, is that we help them solve their very toughest digital problems. And so that tends not to be coming up with a super creative campaign. That tends to be for us, figuring out how to bring all of the data together in a modeling environment that actually creates insights and reports that are actionable, how to create the pipes that enable us to feed that in, either through the insights in an analyst or through a data feed into a platform, how to build web experiences that drive conversion. But I don’t have an answer to the creative side of things. I don’t know that anyone does. 

CJ: Listen, my perspective on the agency of the future… I think that the agencies, you know, in the days of Mad Men were used to holding a lot of power and control, whether that was negotiating really competitive pricing with the large publishing houses, the large TV networks, the large magazines, large everything else, whether that was being so much viewed as the expert. That kind of old-school belief of like, “Yes, I have other agencies that I’m competing with, but I am able to hold my own in the room just because of the environment at the time” is one that I think the agency of the future really needs to completely shed. 

Because the power is in the hands of the consumers, and the power is in the hands of the buyer. I think there are brands who for years have felt like they needed to have more innovation with their agency partner. They needed to have a better client experience. They needed to have a tech-enabled agency. They needed to have an agency that cared about them and was committed to their success. 

So I think the agencies that really thrive today are incredibly client-centric, outcomes-oriented, easy to work with, transparent, trustable. I think there are a lot of those factors that come in, and of course, innovative and creative and bringing a lot of the innate tools to bear. But there’s a little bit of a mindset shift I think that is needed for the agencies that are going to compete and win in the future. 

Aaron: I agree. And I think by extension or by implication, agencies that play well with others. So the sharp elbows, and “I’m trying to claw business from you” model that used to exist is not one that we ever employed, at sometimes very much to our disadvantage when we encountered it in others. But I think we succeed when we’re very collaborative and, as you said, very open.  

CJ: And that by definition is client-centric, right? Consulting as a service industry, being an agency is a service industry, and understanding that you are there to help your client get outsized results and a high degree of confidence that they can drive their business forward in the way they need to is the thing that will help humans be relevant, not fully replaced by Gen AI or AI. And it also means partnering with others and collaborating well on behalf of the client, not on behalf of your own interests. 

Aaron: As I listen to you, I do realize that I do have a perspective on creative, kind of a heretical perspective. And it is that the value of the pitch and the concept is, to me, significantly diminished, because the initial germ of the idea and the concept has value. But as I alluded to, if creative is the new optimization, it’s the variations on that pitch, it’s the variations on the concept and the extension of that concept in a way that tunes it to drive performance that really delivers business value. You can still be true to the original notion, but that notion is only the first relatively small step. There’s a lot that comes after that, in a way that I think hasn’t been so true in the past. 

CJ: I agree with that, and I think looking at the idea content, the creative all the way through to the distribution method, if you’re talking about optimizing creative for a search campaign versus a mobile display ad, versus all these other different form factors, understanding the time and space that it’s being digested and what’s going to be most effective is also really closely linked to that. And not being lazy about it—you need to make that creative excellent in every version of itself. 
 
Aaron: Actively, every day, all the time. 

CJ: My least favorite ad of all time was one that blocked the login page for my fantasy football league, and it was hysterical because it won an award for the best performing ad at a conference I went to later that year. And I’m like, “That stupid ad got so many clicks because on mobile devices, it expanded out, and you literally couldn’t log into your fantasy football team to start your lineup without clicking on this dumb ad.” So I do think rooting out false signals is always a good thing as well. 

Aaron: My least favorite ad right now is one that I just saw as part of the Olympics coverage, and it’s a Gemini ad that suggests that you should write a complimentary inspirational letter to your daughter using AI.  

CJ: That kind of goes back to my point on email. I mean, I know I need to use it to help me more, but I also feel like for those really meaningful personal moments, I want to do it myself. 

Aaron: If you don’t have time to express a sincere and authentic sentiment to your daughter, it sort of calls into question your basic humanity. 

CJ: But would she rather have a Gen AI letter or no letter at all? 

Aaron: I suppose, I suppose. I’m still not happy with the ad. 

CJ: Far more noble source of complaint than mine of “I’m irritated I can’t quickly log into my fantasy football league”. It’s a values-based decision versus an annoyance based judgment. 

Aaron: Both valid though, both valid. This was fun. I appreciate it. I enjoyed talking with you. 

CJ: Me as well. Thanks so much for having me in. 

Aaron: Yeah, thank you. 

Questions or comments

Please let us know if you have questions or comments about this episode by emailing Grace Johnson at grace@wheelhousedmg.com. Want to be a guest on a future episode? Fill out the Be a Guest form at the top of the Digital Clinic page to submit your inquiry.

Make sure to subscribe to the Digital Clinic on Spotify, Apple Podcasts, or Amazon Music to stay up to date on our weekly episodes.