When it comes to embedding AI into your business, taking technology out of the equation is – somewhat counter-intuitively – the best starting point. Meanwhile, agencies are in an enviable position and should realise that they are sitting on a “data goldmine” that’s highly attractive to AI models. That was according to a panel of AI and market research experts speaking at Agency Owners & Leaders 2025 ( 20th November).

The event kicked off with chair Alison Bainbridge, founder of Kokoro Global, painting a somewhat bleak picture of 2025.

Noting the challenges faced by the UK, Bainbridge cited a cash-strapped government seemingly beset by inaction, ailing consumer confidence, slashed client budgets and the challenges of recruitment – both junior and senior.

The third thing that keeps her up at night, she said, is AI. 

“As an agency we’ve really tried to be on the front foot with it, but there’s a sense of storminess,” she told delegates. “It’s really hard to know if you're doing the right thing around AI. 

“I think we're slowly, kind of getting our arms around it, working out how to use it, but it can be a lot of investment and a lot of blind alleys.”

Her intro was a cue for a light to be shone on some of those dark corners, with the ensuring deep dive into the topic overseen by Rebecca Cole, CEO at Cobalt Sky.

“Humans and AI: reimagining agency operating models for the future,” opened with Cole noting how AI has evolved from something talked of hypothetically to a factor that’s “reshaping how our businesses operate, from how we design our research and analyse our data, to how we think about creativity and talent”.

The big question, she added, is: “How do agencies like yours evolve your operating models, not just to keep up but to unlock genuine growth and differentiation in an AI-driven world?”

There would be two aspects to the discussion – developing an AI integration strategy and navigating the human side, the impact of AI on people.

James Grant, co-founder of consulting firm Iwantmore.ai started by pointing out the sheer pace of AI development.

“There’s not a single day that goes by without loads of news stories about something happening in AI,” he said, adding that that “made it almost impossible, even if you work in the industry, to keep up with what’s going on”.

The answer? Taking a step back in order to assess where AI can be deployed to help what a business does for a living, in this case market research.

“When we speak to businesses, technology is completely taken out of the equation at the very beginning of the process, and we try to understand where your use cases are across your entire business.”

Matthew Lintern, global chief executive at MMR Research Worldwide, explained how it has taken a year-and-a-half to avoid a tendency to “jump in and try to digitise everything” and instead to do as Grant suggested – bringing in an outside perspective to look properly at processes.

His agency’s focus has been twofold: “ One is, if a computer can do it, it should do it… it’s about trying to change a mindset.

“I think that can lead to some of the problems you’ve got where you scurry to do everything. But it’s been a challenge for a lot of our people, because they don’t want to think the computer can do what they can. And the extent to which that is true has changed fundamentally in the last two years.”

Secondly, now that his agency is developing AI more strategically, “we’re starting to try and think about margin a bit more”.

“At the moment, it’s getting harder and harder to make money, and I think my one worry with a lot of the AI that we’re doing is that the clients are extremely good at buying, and they’re going to squeeze the hell out of us, and we’re going to end up giving up all that value back.”

This has led MMR to find things internally that “allow us to be better at what we do, not just be more efficient… so we can defend that premium”.

Data goldmine

Steve Phillips, founder and chief innovation officer at consumer insights specialist Zappi, concurred that agencies should consider AI an “efficiency play”. 

But, he added: “I think the much more interesting space, particularly for us as an industry, and the reason I think we’re more advanced in AI than, say, marketing, is that we’re very good at data. And these models are very hungry for data. Not only are we good at data, we have great data; the data you’re sitting on is a goldmine.”

Using video ethnography as an example, he suggested that the AI models will be snapping up consumer insight companies “because they need that data”.

“What we should be thinking of as individual agencies and as an industry is the power of that data and what extra we can do with it, what different things, what different creations we can create with this data.”

“This revolution needs data. And we have it.”

The human side

The panelists were volubly enthused about the various applications of AI, but a key component had so far been missing from the discussion.

“We all know that AI can bring both excitement and anxiety to teams,” Cole noted, encapsulating the ambivalence felt by many.

She asked Lintern how his teams at MMR had reacted when AI was introduced.

“We're dealing with human nature, right?” Lintern said. “A year ago, we asked people how important they thought AI was and how exciting, how urgent they thought it was. Everyone thought it was important. Only half of them thought it was urgent.” But he acknowledged that people are worried about their jobs.

A means of alleviating that worry, Zappi’s Phillips stressed, is separating facts and fears when communicating about AI internally.

“We said to the board and then we said to the whole company that we’re ‘AI everywhere’, so we went through a process of looking at ever function with them, we’re still doing that internally and externally.

“But we also said that AI should enable us to double revenues without changing headcount.

“Part of the reason for saying that was partly to say we’re very ambitious about this and it was also to say no one is going to lose their job.

“Now we may have to change [people’s roles] because [if] they’re reluctant to use it, they’re not going to become more efficient. But we will keep the same headcount and we’ll double revenues.

“I think that worked as a message because it relaxed the ‘I might lose my job’ [viewpoint] and it also made [them] think ‘I’ve really got to get behind AI because that’s the direction we’re going’.”

Training and up-skilling

Getting on the right track clearly warrants significant investment and a shift in mindset, all of which is dependent on training. Accordingly, Cole asked Grant what areas have proven most valuable to employers around training and up-skilling.

He said that businesses often make a decision on a tool, for example Microsoft’s Copilot, deploy it and hope for the best.

“‘That’s usually what happens: ‘Oh it’s intuitive, I’ve just used it, everyone will just pick it up.’ I mean, that’s just nonsense, it doesn’t happen. But even just running bog standard training courses doesn’t help either. 

“The only way we've got it to sink in is when we build training courses that are contextual to what that individual does for a living. Until you get to the point where you’re showing them specifics of how functionality of tool A or B helps them in their job, it’s not gonna land.

“So just saying, watch this YouTube video or going to this ‘lunch and learn’ on Copilot, or whatever tool, ain’t gonna work.

“It’s got to be: Bob does this, and you show Bob how this can help them in their spreadsheet, or whatever it is, so it’ll be contextual to them as an individual.

“That takes a lot of work, and that’s either expensive by bringing people in, or it takes a lot of resource internally to build that and do that, but it is the only way you’ll get it to land.”

Cole concluded things, asking the panel to each share a key takeaway.

“Pile in and play,” said Grant, a sentiment shared by his fellow panellists, who expressed it slightly differently.

“If you don’t do it, you’re not going to get there. You cannot sit and wait. It’s going to be too late,” said Lintern.

Phillips added: “AI literacy for everyone in your organisation regardless of role.”

The human factor

The subject of AI came up elsewhere during the day’s conference. In a session about the rise and fall of research business Flamingo, its co-founders and former joint-CEOs were asked to share their thoughts on how to build a successful business in the current market. Both cited AI.

Maggie Collier offered some rather reassuring words, a balm to those worried generative AI is going to make people redundant: Imagination and storytelling are going to be more relevant in the age of AI, she said.