When AI is combined with first-party data, its power multiplies, explains Esme Robinson of Epsilon. Get ready to boost your brand.
Successful marketers build systems that learn from real behavior and improve with every outcome, explains Robinson
Artificial intelligence in marketing is at fever pitch. Titans like McKinsey have lent credibility to the hype, stating companies that deploy AI are already recording revenue lifts of up to 15% and sales return gains of 10-20%.
That momentum is fueling big promises, like Meta claiming that by 2026, its platforms will be able to plan, create, and place an advert end-to-end with almost no human input. The headline is striking, yet it glosses over an awkward truth: the gains being reported today come from selective, supervised use of AI rather than fully driverless systems.
If we speed blindly toward complete automation, cracks are likely to appear in any model still reliant on incomplete or inaccurate data, or shallow identifiers such as emails and device IDs. Long-term growth needs a richer foundation.
AI must combine those basic elements with consented behavioral and transactional data that explain what people actually do, not just what they click. Only with that deeper view of actions and connected identity can automation keep finding new, real value.
Identity is the missing link. Because when AI is guided solely by limited behavioral inputs, it runs the risk of reinforcing yesterday’s preferences. Keep showing the same fast-fashion handbag to a customer because their clicks yesterday suggested ‘trend shopper’ and the model never discovers that the same shopper is shifting toward sustainable labels.
Without a richer understanding of the person behind the data – like what they’re buying, what they value, and what influences them today – AI ends up optimizing against incomplete information. We’ve all experienced the echo chamber of stale, repetitive, or simply irrelevant ads that proliferates as a result.
Transactional first-party data is essential to achieving this understanding of shopper identity. Retail media has already set the pace. In these environments, purchase behavior from online and in-store can be linked with real individuals. Loyalty schemes combine consented profiles with basket data, so outcome-based AI can learn which messages move tea, toys, or trainers.
When identity is tied to outcome data, and that information is accessible across web, mobile, streaming, and in-store screens, AI becomes a genuinely useful decision-making engine. That’s when you really start to see success. It’s no longer just guessing what might work; it’s instead evolving with the individual, and staying relevant, without getting stuck in a feedback loop. As with all marketing strategies, AI should put the customer at the center to reap the most benefits.
When marketers connect identity, first-party data, and AI, they keep hold of a crucial layer of control, rather than putting everything into a sealed box. Generative tools should work in tandem with the foundations the marketer sets, shaping content around the individual, not just the product. The same shopper who bought vegan trainers after watching a film on mass-market fashion can now receive fresh creative that pairs Stella McCartney with a local indie designer, instead of yet another generic fashion ad.
Another important advantage of working within a retailer’s ecosystem is the ability to manage how the brand is presented. Marketers aren’t forced to give up creative control to an AI system that sees only part of the picture. Instead, they have the opportunity to shape how the brand looks and sounds in each setting, harmonizing tone, design, and overall feel to ensure consistency, no matter the channel.
This level of influence matters. There’s a clear difference between how a brand appears alongside search ads and how it appears on a premium retail platform like John Lewis. One is engineered to feel high-quality and aspirational. The other prioritizes standardization and reach. Within the retail environment, marketers can align the brand with more considered experiences rather than letting it become just another generic result.
Here’s a simple checklist for success:
Full automation alone won’t deliver results; it still needs human insight and solid data to guide it.
If you don’t have a lasting understanding of your customer’s identity, AI will fall into a looping pattern.
Stay present in the process and keep control over how your brand looks, sounds, and evolves across every platform.
The real value of AI is in helping you move alongside your customer, staying relevant in real time, at scale.
The marketers making a success of AI aren’t chasing scale for scale’s sake; they’re building systems that learn from real behavior and improve with every outcome. That takes a mindset shift. The future isn’t a handover to machines, it’s a partnership where AI does the heavy lifting and marketers stay close enough to the data to shape what comes next.
Epsilon
Epsilon is a global advertising and marketing technology company positioned at the center of Publicis Groupe. We connect advertisers with consumers to drive performance...
