The question on many marketers’ minds at POSSIBLE 2026 this week wasn’t whether to integrate AI, but rather what’s working, what isn’t and what are the real-world strategies and frameworks they can emulate to achieve results.

If the answers from some of the world’s most recognizable brands are any indication, there’s no single playbook. But there are lots of lessons learned and, after a period of sufficient experimentation, clear examples of business impact.

At Walmart, for instance, 73% of the company’s marketing investment is now AI-enabled in some form — touching targeting, bidding, media placement and dynamic creative. At The Home Depot, the marketing team has carved out dedicated “Transformation Thursdays” to map end-to-end workflows and identify where AI can streamline them.

At Coca-Cola, what began as a handful of generative AI experiments has evolved into a commitment to embed AI across the entire marketing process — from consumer insights to creative to media buying to performance measurement. And at U.S. Bank, the team built — and continues to build — synthetic consumer personas from over 1,600 curated data sources, capable of reacting to competitors’ ads and breaking news in real time — all in a heavily regulated industry.

Following are insights from the POSSIBLE stage from top leaders at The Home Depot, Cloudflare, Walmart, U.S. Bank, Coca-Cola and PepsiCo on how their brands are approaching AI integration, implementation, adoption and agility within their marketing organizations.

Molly Batin, CMO at Home Depot, on how The Home Depot structured its internal AI adoption process:

“From a marketing team perspective, we [said] let’s map out our end-to-end workflows. And honestly, that was enlightening in itself, of how complex we have made our own processes. So how do we simplify our processes, get ourselves out of our own way so that we can then streamline and automate?

We have set aside Thursdays as Transformation Thursdays within the marketing organization, and really leaning into okay, as we think across the entire end-to-end workflow, where can AI plug in and truly help us solve challenges, work faster, be more efficient, be more effective, and really taking this more strategic approach as opposed to, we’re just going to plug holes and try do a whack-a-mole approach …

[We’re] really thinking about, where are our biggest pain points? Where can we move the needle the fastest, to see those returns and then get that flywheel moving? So that’s how we’ve focused on it and we’re now doing a lots of pilots and we’re moving from pilot to more automation at scale.”

Matthew Prince, Co-Founder & Chief Executive Officer at Cloudflare, on the three practical steps every marketing organization should take right now to prepare for AI:

“The first thing everyone should be doing right now is looking at their organization and saying, sure we’ve got these tools and applying them to the same frameworks and same structures probably isn’t going to give you that one giant thing. You’re not going to get these passive productivity gains. And so what we’re doing is breaking down across the entire organization, but in marketing in particular, and saying, what are the different ways that, if we have these new tools, we’re going to be able to design this?

I think the second thing is that you’ve got to be asking yourself, am I really adopting this? There’s so many of us that kind of toy with it or play with it, but they don’t bring it into their actual real day-to-day work. And I am trying really hard to say, ‘I’m going to get back to this.” And so when I look at something that’s broken on our site, where in the past I would just kind of note to someone, ‘Hey, this is broken on our site,’ I’m now going in the code and working with tools to actually rewrite that feature and then bring it back to our team saying, ‘Hey, here’s this thing, here’s a fix’ and bringing it in.

And then I think the third thing is just really trying to learn as much as you can, looking across the industry and just saying, what’s changed? … The business model of the internet is going to fundamentally change how we all do our jobs. Being open to what that change is going to look like is probably the most important thing that we’re doing.”

Michael Lacorazza, Executive Vice President & Chief Marketing Officer at U.S. Bank, on the most important piece of advice for marketers building AI-powered synthetic personas or any AI research capability:

“If you’re interested in building an ecosystem like this, don’t start with a large language model technology. Start with the data sets that you think most distinctly represent the profiles that you want to create and how they might influence many decisions. Because once you get that part right, you’re kind of set up and firing on all cylinders. From there, it’s actually a pretty simple process to layer on the LLM.”

We also started with having a very contained set of experiments that were low risk, that people could react to and see. And we have also done a fair amount of validation with live audiences answering some of the same types of questions. So we to get a kind of human comparison to the work, and we continue to benchmark that along the way.

The thing is, just because you’re getting an answer from a person doesn’t mean it’s a great answer — there are so many flaws and research biases, things like people not telling the truth. How do you trust that just because it’s a real, alive human … it’s necessarily a better answer?”

William White, Chief Marketing Officer at Walmart, on the company’s philosophy for integrating AI while keeping people at the center:

“I think it’s important to recognize that being AI-first doesn’t mean people-last. And we as marketing organizations really need to understand how we enable the humans on our team to do the best work and use the technology to help power some of that — but it’s never going to replace that …

I can tell you that every pocket, every sub-function within the marketing team is leveraging AI in the work that they do today, and absolutely embracing it. Today, 73% of our marketing investment is in some way AI-enabled, whether that’s in targeting, whether that’s in our bidding optimization, media placement, whether that’s in dynamic creative — it is dramatically making an impact on our business.

I think, if anything, the push where we need to do more is that this is happening in each team, but the wiring and the connection, the workflow, isn’t necessarily AI-enabled in a way that it could or should be. And so that is a little bit clunky for the team right now … There are still handoffs that feel overly manual. The friction is much more about how do we move faster and get things connected …

We’re the largest employer in the country with 1.4 million associates, and anyone’s opinion of Walmart is probably shaped more than anything else by the interaction they’ve had with an associate in our store. So recognizing that people and humanity are the drivers of our business is really important.”

Manuel Arroyo, EVP, Chief Marketing Officer & Chief Commercial Officer, The Coca-Cola Company, on how Coca-Cola moved from early AI experiments to embedding AI across the full marketing enterprise:

“We started the journey with a few use cases, a few experiments, a lot focused much more on generative AI basically — and that very quickly gave a clear indication to a few of us that we could see AI being embedded into the whole marketing end-to-end process, all the way from insights to creation, media planning, media buying, performance tracking, and into all the workflows that connect the whole enterprise.

So we’ve been working on that journey now. The one I’m really excited about in particular is creative — what we’re very, very clear on is that AI will always be a tool — or at least for the foreseeable future — for a human storyteller. If you don’t have phenomenal human storytellers crafting the right level of creativity, bear in mind that … we don’t sell only liquids — we try to sell emotions, convey emotions. For that you need a human that crafts that very carefully.

Now, AI is allowing all of them to do that at a much faster pace, at a scale that was completely unprecedented, and to levels of customization that were unthinkable. For example, the World Cup original video right now has more than 160 different combinations out there in the world. And then when it goes into media that spans 100 clusters across many parts of the world We are communicating seven or eight very distinct clusters with adaptive creative — and we’re doing that on the fly.”

Mark Kirkham, Chief Marketing Officer at PepsiCo Beverages North America, speaking about an exclusive Walmart Connect product launch and what the new model of flexible, test-first marketing looks like in practice:

“The days of just launching with one big spray-and-pray national launch — I’m not saying we’re not going to do that, we’re PepsiCo, we still drive innovation at scale — but as an industry, our ability to be flexible by channel, working with partners early to understand, because most of our partners are driving growth through omnichannel sales … The only way you’re going to be successful is to be flexible. Flexible in how you media plan, flexible in terms of your launch strategy, flexible in terms of your messaging strategy. And by the way, we used to not be that. We were highly linear.

I think the other thing is that the funnel — it doesn’t exist like that anymore. It is a messy map of sorts … And so this linear nature of brand building, this linear nature of media, has changed dramatically over the last few years. Being flexible, thinking earlier about innovation, being able to dynamically change your message — that’s what’s making a difference for the consumer.

Traditional retail relationships can be hard … But as marketers and as media partners, we need to share more data. We need to have more conversations early. We need to lean in on innovation launches on your channels and unique channels and sometimes exclusive channels. And that can only happen if we’re not just talking to merchants and we’re not just talking to marketing people and not just talking to media people. We need to bring everyone together … When you have those insights and you can learn what’s working and what’s not, being able to adapt, and when you have a powerful network like Walmart Connect, you can adapt — and that will immediately result in sales.”

Editor’s Note: More coverage from POSSIBLE is forthcoming and includes exclusive video interviews with top CMOs attending the conference.