The point when Sorin Patilinet realised the marketing profession had started to get serious about meaningfully integrating AI into workflows was, he jokes, “when people stopped just making animated characters”. The public release of ChatGPT in 2022 prompted introspection and plenty of rumination, but mostly, he adds, it turned marketers into amateur animators determined to create humanistic animals and use them in thrown-together campaigns for everything from credit cards to chocolate spread.
Four years on, the cute bunnies are mostly gone. LLMs are powering not just chatbots but multi-faceted, widely deployed tools. And Patilinet, head of marketing effectiveness and measurement innovation at PepsiCo, is keen for a more mature debate about what it all means for marketing practice.
A measured relationship with AI, he says, begins with deciding what’s necessary and helpful rather than tokenistic or gimmicky. “The promise of today’s generative AI is having everything at your fingertips. You can create whatever product you have in mind. You can create a new flavour of drink, design the packaging, create the brand and everything is done in 10 seconds.
“But that’s where the challenge comes: just because you can do it doesn’t mean you should. Just because you can create a mango passionfruit flavoured cola doesn’t mean you should invest your entire supply chain capacity, your sales force and put your entire company behind an idea. There’s still a role for understanding markets, category dynamics and consumer research.”
He isn’t the only thinker out there looking to temper AI hype with reality, but he brings a very specific lens. The Romanian-born lecturer, author and practitioner cut his teeth in the profession at British American Tobacco before 13 years at Mars, where he pioneered an effectiveness-driven approach to analytics and measurement, which culminated in the release of his book Marketing Effectiveness last year.
It has put him at the forefront of a movement championing rationality and aiming to take deep data analysis from specialism to centre stage. As he puts it: “As marketers, we should measure, discuss and question all our decisions through the lens of the final outcome. All the strategy, pricing, product, media, creative, measurement – they shouldn’t be left to chance, they should be done because of a particular business outcome.”
What this means in practice, Patilinet will discuss in a keynote at Campaign360 in Singapore in May. His key assertion is that AI-driven market research is a killer app for marketers, and the raw materials are now in place for a data-led revolution, particularly among consumer businesses. “We have consumer panels that we can run experiments with. We have extensive data sets of loyalty card data from retailers that enable us to know exactly how duplication of purchase works, how people are buying into different categories, where we have gaps and where we can potentially launch a new proposition.”
The next stage, adds Patilinet, is to “get more quality into the data”, which potentially allows for better-informed decisions on budget allocation. “If you talk about sponsorships, you can now understand exactly the number of segments in which your logo was visible in different video formats, and the number of pixels that were available to you during a live sports broadcast. That improves the quality of the model. It runs almost instantly, so as a marketing analytics function, we’re able to make recommendations instantaneously.”
Human judgment still sits at the end of these processes – for now. Patilinet says he has already had a “great experience” using synthetic data, which creates composites of an audience from available information, in consumer research, and the technology enabling accurate digital twinning of individual consumers is advancing rapidly. He believes this has particular resonance in B2B markets: “Good luck interviewing 25 CEOs of large banks if you want to build a product for them. But you can build personas of them.”
At Campaign360, Patilinet will present these ideas among a wider exploration of the “human edge in an intelligent world”, drilling deeper into the role of creativity – still synonymous, within marketing, with humanity – in an age when AI is making more possible, faster. And he’ll do so with reference to Asia Pacific, where he says there is “an openness to information and insight, and less endless debate” among marketing leaders, and where data is less of a bottleneck to decision making.
But the central quandary remains whether we will use AI to do the same things more efficiently or work smarter. That isn’t a straightforward equation. At a basic level, AI “allows us to focus on strategy and big thinking… which is great if you need that much strategy.” Undoubtedly, there will be a need to reconsider staffing and supplier relationships in the coming years. “Marketers of the future… will still rely on a multitude of friends from agencies, research agencies, thought partners and media partners,” says Patilinet. “But relationships will be more fragmented than today because you will dip in and out depending on need.”
It means discussion of the role of the individual in marketing’s future is more than theoretical but will define the structures and outputs of tomorrow and the skillsets that are (and aren’t) irreplaceable. A Campaign study in late 2025, for example, found leaders were thinking deeply about these issues, even if the widespread uptake of AI tools had not yet translated into significant time or efficiency savings.
Patilinet agrees that asking how marketing has adopted AI is “like asking how we’re adopting electricity in the age of fire” but argues that’s because we’ve been too busy focusing on the next big thing rather than appreciating what the technology can do in the here and now. “As a profession we went very quickly from ‘we have an amazing solution’ to ‘let’s crack creativity’.
“There’s so many things that AI can help us with – workflows that are now twice as fast, briefs that are better written, all this back and forth between us and creative agencies is happening a lot faster. But sometimes we are a little bit obsessed with using AI to come up with the next 1984 from Apple or Just Do It from Nike, when in fact that’s the hardest part of our role.”
Where AI should add value next, he adds, is “decoding the value of business outcomes” by helping marketers understand the effect of their work on sales, brand equity or other key metrics. For the past nine months, Patilinet has been trying to put just such ideas in place at PepsiCo, initiating a “fundamental shift from pre-testing to in-flight testing”.
“I’ve seen a massive shift from ‘let’s spend a lot of time creating ads and then testing with consumers so we are ready to go live’ to a phase, enabled by technology, in which there are multiple executions, you use them on social and digital platforms where you can collect data and the winner stays there and the loser gets withdrawn.”
It comes at a time when FMCG marketers are at the sharp end of AI integration, with Mondelez touting cost savings of up to 50 per cent in content generation and claiming it may produce a 2027 Super Bowl ad using the technology. PepsiCo says it wants to be “agentic in every part of the business” this year but has been more circumspect in its use of AI for creative to date.
That’s in keeping with Patilnet’s belief that, in lieu of a proper understanding of how we should deploy AI beyond the transactional, its use in creative has been “a little bit vanilla”. He adds: “We’ve done a lot of models that predict through AI the behaviour of consumers when they see advertising and what struck me looking at that, is the response of the model correlates with the response of real people but the variation in that response is much smaller.
“So if you’re putting 1,000 ads through consumer research… when you put it through an AI predictive model that mimics consumer response, it narrows to a smaller band. Everything becomes average and it’s harder to make decisions.
What he’s looking for, he says, is “creativity in service of a business outcome”. But refusing to embrace AI in the meantime “isn’t an option… If it doesn’t work today, it will work tomorrow. What can you do as a marketer to get ready for that future in which you will have to work hand in hand with AI?” Patilinet doesn’t have all the answers, but he knows which team he’s on: “I’m rooting for the humans.”
| Sorin Patilinet, global marketing effectiveness lead at PepsiCo, will deliver the keynote at Campaign360 in Singapore on May 20-21. |
|---|
