Three recent U-turns I’ve made about AI and the future of healthcare insight

By Abigail Stuart, Founding Partner, Day One Strategy

I’ve been working with someone recently who has really challenged my thinking. One of the best conversations we’ve had was about the things we’ve got wrong, and why it matters to admit when our thinking has changed.

It made me reflect on AI and the future of healthcare insight because, over the last couple of years, I’ve changed my mind about several things.

Here are three U-turns I’ve made.

U-turn 1: I thought having access to the most capable AI was the competitive advantage

Like many people, I initially assumed the competitive advantage would come from having access to the best AI. Which model should we use? Which platform is most capable? How much of the research process can we automate?

These are important questions, but I no longer think they are the questions that will define the winners over the next five years. Today’s mainstream models, whether that is ChatGPT, Claude or Gemini, are already powerful enough to change how we work.

Access to powerful AI will not remain a meaningful point of difference for long. It will soon become table stakes. Every company will have access to technology that can analyse data, identify patterns, summarise findings and generate outputs. In the same way we do not choose partners because they use Microsoft PowerPoint or Teams, I do not think clients will choose partners simply because they use AI.

The real advantage will come from what organisations build around AI. That means the quality of the data, the depth of the customer understanding, the commercial judgement and ability to connect all of that to the decisions clients need to make at the moment that matters.

U-turn 2: I thought primary market research might become less important

At one point, I wondered whether AI would reduce the need for primary market research. If you can test ideas with synthetic respondents in real-time, at a fraction of the cost, surely clients would seize the opportunity to move away from expensive, slow market research.

I still think lower-importance questions and decisions will increasingly be answered using existing knowledge, AI-assisted analysis and synthetic approaches. Some projects that would previously have gone straight into research may not need to.

But I now believe the more important AI becomes in facilitating decision-making, the more valuable high-quality primary research becomes. AI is only ever as good as the customer understanding it learns from. If the underlying evidence is weak, outdated or biased, the outputs will be weak, however impressive they sound.

That means research that uncovers real customer behaviour, emotional drivers, unmet needs, decision-making tensions and barriers to change becomes more important, not less.

Increasingly, I think that primary market research will be seen as one of the most valuable strategic assets a healthcare organisation owns. Not because it answers one question, but because it can power many future decisions.

U-turn 3. I thought about research primarily as a series of projects

Looking back, I realise I’d largely thought about research as a series of projects. We don’t ignore what we’ve learnt before, far from it. But connecting previous primary research, competitive intelligence, market evidence and internal knowledge has always taken significant time and effort, which means we haven’t always been able to do it as consistently as we’d like.

I now think that’s where we’ve been limiting ourselves.

The challenge has never been recognising the value of what we already know. It’s been making that knowledge readily accessible at the moment a commercial decision needs to be made. Until recently, that simply wasn’t practical.

AI changes that.

That changes the starting point. Instead of commissioning new research as the default response, we can first fully interrogate everything we already know to understand what evidence already exists, what assumptions need challenging and where the genuine knowledge gaps lie.

That fundamentally changes the conversation. Instead of asking, “What research do we need?”, we can ask, “What decision are we trying to make, what do we already know, and where are the genuine evidence gaps?”

Sometimes the answer will be new primary market research. Sometimes it will be making better use of the intelligence that already exists. More often, it will be a combination of both.

That’s what makes continuous intelligence possible. Not an archive of old reports, but customer understanding that grows over time, making every investment in research more valuable than the last and giving teams greater confidence in the decisions they make.

What’s next for AI and healthcare insights?

I don’t have a crystal ball – AI is moving too fast for anyone to accurately predict what the future holds. I’m certain I will continue to adjust my views and perspectives and more U-turns are inevitable.

But I do think the future of healthcare insight won’t be defined by who has access to the best AI. It will be defined by who is best at meeting the decision-making needs of clients, in a timely manner, by combining human understanding, primary market research and AI.

If that’s true, then our job is to ensure every investment in customer understanding continues creating value long after the original project has finished.

I suspect my thinking will continue to evolve – I hope it does!. Because if AI is teaching me anything, it’s that the most valuable thing we can do isn’t pretend we’ve got all the answers. It’s stay curious enough to keep asking better questions.