Stop building faster horses: what AI should really change in market research

By Abigail Stuart, Founding Partner, Day One Strategy

I’ve been watching the industry talk about AI for long enough now to see the pattern. Most of what’s happening is optimisation rather than reinvention. We’re using powerful technology to shave minutes off processes we know aren’t that valuable to begin with.

It might be useful, but it isn’t driving the transformation we’ve all been promised.

The problem is that we’ve fallen into the trap of trying to protect the way research has always been done instead of asking what research is actually for. And if we don’t deal with that head-on, we leave ourselves wide open for smart outsiders, without our baggage and “best practices”, to disrupt the sector entirely. If we’re not careful, we might end up the Blockbuster of insights, leaving the Netflix’s of the world to rewrite the rules.

If you strip research back to its core, the job has always been simple:

• Help organisations understand people
• Help them make better decisions

Everything else we’ve built around that purpose – the sample frame, the discussion guide, the tracker wave – is scaffolding. And somewhere along the way, we started treating the scaffolding as sacred.

You see it in the way we obsess over whether a Likert scale should be 5 or 7 points, as though the world hinges on it. When was the last time you asked a colleague to rate their weekend on a scale of 1–7?

You see it when we proudly present 60-page debriefs even though the client only needed the one slide that actually answers the question.

These things aren’t meaningless, but they’re just ways of getting closer to how the wider market thinks and behaves. The participants themselves aren’t the goal, they’re the lens.

AI used to replicate old processes is a waste of everyone’s time

A lot of what’s being sold as innovation today is simply AI photocopying what we already do:

• AI moderators acting like human moderators
• Synthetic respondents behaving like survey respondents
• AI that writes the same questionnaires we’ve always written

It’s clever, but it isn’t very ambitious. And worse, it often lands badly because it feels like replacing human skill rather than elevating it.

If we stop treating traditional research as the blueprint, the innovation roadmap changes completely.

1. AI can give us a better picture of reality than fieldwork ever could

This isn’t about AI pretending to be a respondent. It’s about bringing together signals from everywhere. what people search for, what they buy, what they talk about, how markets move, how culture shifts.

2. AI can help us understand why people behave as they do

Human behaviour isn’t random. It follows patterns shaped by beliefs, habits, pressures, and circumstances. AI is extremely good at connecting those dots because it can see patterns across millions of moments rather than a handful of interviews. Taking to people is still essential, but when we combine this with other data sources it is more powerful.

3. AI lets us explore scenarios, not just reactions

In the traditional model, we ask people to imagine the future, which humans are notoriously bad at. For example, try asking a doctor how patients might react to a new treatment concept. They’ll give you a thoughtful guess. But they can’t simulate all the variables such as real-world pressures, reimbursement, patient fears, cultural norms.

In an AI-enabled model, we can simulate how adoption might play out based on patterns in similar launches, clinical data, economic signals, demographic behaviour, and cultural attitudes. Then we take those scenarios to real people and see how they react. That’s far more powerful than asking doctors to predict the unpredictable.

4. AI can turn research into a living intelligence layer

We won’t be running bursts of research forever. Not when we can have a constant stream of signals forming a living, evolving picture of the market.

Our role shifts from producing insights to interpreting a continuous flow of intelligence. Far more valuable and strategic.

5. Humans become more important, not less

Once AI removes the noise, the human contribution becomes more essential: Judgement, challenge, context, understanding and strategy.

If AI frees us from clinging to old methods, we can get back to the thing our clients actually need:

• A better read on human behaviour
• A way to cut through uncertainty
• Clearer direction
• Faster, more confident decisions
• Less guesswork

In other words, research becomes decision intelligence – the engine that helps senior teams act with clarity.

The risk? We cling to the old model because it feels safe

If the industry keeps using AI to do what we already do, just faster, research will become cheaper, more automated and less central to decision-making.

But if we decide the purpose comes first and the method follows, AI becomes a genuine accelerant and amplifies our impact.

That’s the choice in front of us. Faster horses or a new way of thinking entirely.