If there’s one department that never met a shiny new technology it didn’t want to test-drive, it’s marketing. Over half of brand marketers are already using AI for content creation, according to recent research from Marketing Week. But behind that adoption rate lies a more nuanced story: the difference between using AI and using it intelligently enough to live up to the hype.
At a recent panel discussion, three marketing leaders shared their own experiences with AI’s rapid evolution and their advice for other businesses.
AI’s evolving role in marketing
Lisa Harkness, partner at McKinsey & Company, opened the discussion by positioning marketing as “the bellwether of the company in using data to generate value with customers.” She noted that a lot of GenAI’s total potential value would come from marketing. “We estimate that GenAI could contribute up to $4.4tn in global productivity. And marketing and sales is one of four groups that combined could reap an estimated 75% of that value.”
David Pugh-Jones, CMO of the research engine Corpora.ai, echoed this sentiment: “Marketers are always going to be the guinea pigs when any new technology appears,” he noted. “We’re the ones that have to dive deeper and work out the true value and then eloquently explain how that can be transposed into a business framework.” He added, “I don’t think it’s about if we use AI, it’s about how intelligently we deploy it.”
Agentic AI is the latest iteration of the technology on marketers’ radar. Lauren Maynard, global chief growth officer at FutureBrand, explained that agentic AI is much closer to the revenue line than generative AI because it can power clear solutions. She said, “It’s a lot easier to frame those as viable solutions. That makes it easier for people to buy them, consume them and implement them within their business. That’s really important for how we’re able to talk about it with clients.”
Getting an AI strategy right
For all the excitement about AI’s potential, implementation must be carefully thought through. The difference between success and expensive failure often comes down to having the right adoption strategy – something FutureBrand has considered carefully.
Maynard provided a framework for how her organisation integrates AI into marketing and creative workflows, focusing on four areas. The first area of consideration is ideation. In this context, that means training LLMs on all of the rich, unstructured data that individual marketers might read, but companies were historically unable to process as a whole. The second is generation, or producing creative ideas much faster than they could be created manually.
The third area is validation, which historically was not readily available in the creative space. Now, marketing teams can use customer data to test a strategy and rapidly evolve it without having to rely solely on external consumer testing. Maynard noted: “You always have to do that at some point, but this allows you to feed back in a way that is concrete and more objective.”
The last crucial area for Maynard is brand governance, although she pointed out that “AI is not the solution. It is a tool that powers better brand governance solutions.” For large multinational organisations, this allows for creative execution that is both locally relevant and globally coherent. “It delivers an efficiency and effectiveness for businesses that they really couldn’t have before,” she added.
This structured approach ensures that creative teams can leverage internal data for inspiration, speed up content development and test ideas before release, all while maintaining brand consistency.
Garbage in, garbage out
Every marketing team has or knows of an AI horror story. Usually, it involves a well-intentioned experiment gone wrong, producing results that range from embarrassingly off-brand to reputationally damaging. The culprit is almost always the same: poor data quality.
You have to be so intentional with the way you structure your queries and map out your end game
“It’s garbage in, garbage out,” Maynard said. “You have to be so intentional with the way you structure your queries and map out your end game.”
Harkness gave an example of a client experimenting with email personalisation. “The first version of the personalised email to a specific segment was: ‘Hey, lovely lady,’” she recalled. “That is not a great way to start.” But she pointed out that this story isn’t an example of failure, but an indication of the importance of having the right guardrails in place to ensure that an early experiment never goes out to a customer. She added that personalisation is something that is seeing great demand, so it’s hugely valuable for organisations to be able to experiment safely with it.
Even with perfect data and well-trained models, AI’s outputs need human scrutiny. “In an investment bank or management consultancy, you would never take the analyst’s first pass analysis and send it to a client,” Maynard pointed out. Yet with AI tools, that’s often exactly what happens. “There’s a trust in tech in this scenario that you don’t have in other spaces – and there’s a lot of danger in that.”
Levelling the playing field like never before
Unlike previous technologies that required specialised teams and substantial budgets, AI tools are increasingly accessible to anyone with a laptop and an internet connection.
“AI is in the hands of the people. It is no longer limited to theoretical use cases or pilot projects,” said Pugh-Jones. “This will level the playing field like we never knew before. It’s going to drive mission-critical functions across every discipline of every business.”
He pointed to AI’s ability to unlock previously inaccessible market intelligence. “We now have the capability to share insight that we never even had access to. We’re unearthing information and data at a scale and speed that was previously unimaginable. The strategic value lies in the technology’s ability to turn the entirety of the world’s data into actionable intelligence.”
However, he noted that people still need to connect to build trust and demonstrate values. “Those things don’t go away. We’re really just looking at how AI helps us to do that”.
Change management and strategic intent
So what should marketing leaders be doing to keep pace with AI advancement? All three panellists agreed: integrating AI is as much about organisational change as it is about technology.
“50% of the effort is in the tech,” said Harkness, “but the other 50% is in the training and reorganisation of how things work. You can’t think you can just buy this super tech that will be immediately trained and everyone will know how to use it. There’s a tonne of change management that goes on that a lot of people don’t see.”
Maynard expanded on this: “You need to create freedom within a framework – and the framework has to come first.” AI tools can only be successful if teams have the room to explore and fail safely. “If it’s too process-oriented, there’s no business in bringing it in.”
Pugh-Jones ended with a reassuring perspective on AI’s role in the workforce. “The future doesn’t belong to AI alone,” he said. “But it does belong to some form of symphony of human endeavour and creativity working in harmony with intelligence. We just need to make sure we’re delivering the message that this isn’t stealing people’s jobs but actually creating them.”
If there's one department that never met a shiny new technology it didn't want to test-drive, it's marketing. Over half of brand marketers are already using AI for content creation, according to recent research from Marketing Week. But behind that adoption rate lies a more nuanced story: the difference between using AI and using it intelligently enough to live up to the hype.
At a recent panel discussion, three marketing leaders shared their own experiences with AI's rapid evolution and their advice for other businesses.
AI's evolving role in marketing
Lisa Harkness, partner at McKinsey & Company, opened the discussion by positioning marketing as “the bellwether of the company in using data to generate value with customers.” She noted that a lot of GenAI's total potential value would come from marketing. “We estimate that GenAI could contribute up to $4.4tn in global productivity. And marketing and sales is one of four groups that combined could reap an estimated 75% of that value.”