And why the next frontier is studying AI as part of the customer journey
Attending the MRS AI: Powering-Up Insights Conference sent a clear and encouraging signal: the conversation around AI in market research has grown up.
Less hype. Fewer promises of magic.
More pragmatism, accountability, and real-world application.
Across brands and agencies alike, AI is no longer framed as a silver bullet, but as a serious, embedded component of how insight is generated, challenged and used. For us at Eumetra, the conference also crystallised a key realisation: while AI as a tool is maturing fast, AI as a phenomenon shaping customer behaviour is still largely underexplored.
Insight is not AI output – it’s decision support
One of the strongest points of alignment across sessions was a reminder the industry badly needed: data and information only become insight when they support decisions.
AI is exceptionally powerful at handling complexity – synthesising vast datasets, surfacing weak signals, accelerating iteration. But the point of insight remains unchanged: enabling better judgement, not outsourcing it. The most credible use cases reinforced this principle clearly.
AI as augmenter, not oracle
Examples shared by organisations such as Shell, Heineken, Co-op and E.ON Next showed AI used at its best:
- challenging assumptions
- speeding up sense-making
- opening new questions rather than closing them
In none of these cases was AI positioned as a replacement for expertise at the moment of interpretation and action. The winning model is AI as challenger, not decision-maker.
From static outputs to living insight systems
Another clear shift: the move away from one-off studies toward living insight ecosystems.
AI-powered personas, simulations and internal insight platforms are becoming standard, helping democratise access to insight while preserving methodological rigour.
This evolution is also redefining the role of insight teams – less about producing answers, more about:
- filtering noise
- shaping better questions
- focusing on higher-value thinking.
Democratisation ≠ loss of rigour
A recurring concern surfaced openly: AI can produce expert-looking outputs without expert judgement. The response isn’t to slow adoption, but to build capability around it – AI literacy, shared prompting standards, credibility checks and clear guardrails.
Rigour doesn’t disappear with democratisation – but it must be designed in.
The question we still ask too little
Despite the progress, one question remains surprisingly under-addressed: are we studying AI only as a research tool – or also as a force reshaping how customers think, decide and choose?
AI is already influencing:
- how information is sought and trusted
- how confidence and authority shift
- how decisions are delegated
- how brands are evaluated and compared
This is where insight meets foresight – and where we believe the next frontier lies.
Eumetra’s complementary approach: studying AI within the customer journey
Building on two years of work with international clients and industry bodies (including MRS and ASSIRM), Eumetra is launching a new research model that treats AI not just as infrastructure, but as an object of study.
Rather than focusing on AI primarily through the lens of efficiency or automation, our approach analyses how AI is already mediating customer journeys – and how this mediation is evolving over time.
At the core of the model is a simple but powerful idea: prompts are behavioural signals.
How customers interrogate AI – what they ask, how they frame choices, what they delegate, and what they retain control over – offers a forward-looking view of emerging needs, trade-offs and decision logics.
What the model captures
The framework focuses on three core dimensions:
- The evolution of AI prompts along the customer journey
Tracking how customers interact with AI across key moments – information seeking, comparison, advice, onboarding and servicing – as an early signal of shifting priorities and expectations. - Trust and decision delegation
Measuring where customers already rely on AI recommendations, where they hesitate, and how authority is rebalanced between AI, human advice and personal judgement. - AI-enabled customer segments
Identifying who is leading behavioural change today – and who is most likely to shape future journeys.
Applied to sectors such as banking, this approach helps anticipate how AI is reshaping discovery, evaluation, brand consideration and switching dynamics – not retrospectively, but as they emerge.
Why this matters now
The MRS conference made one thing clear: the market is moving from experimentation to responsible acceleration. AI is becoming embedded, operational and – finally – discussed with the seriousness it deserves.
The next step is to ensure our insight models evolve accordingly.
AI doesn’t just change how we research customers.
It changes how customers make decisions.
We are currently demoing this new research model with international organisations exploring foresight, innovation and customer strategy in AI-mediated environments.
→ If you’d like to see how tracking AI prompts can reveal the future evolution of customer journeys, we’d be happy to share a short demo and concrete examples of deliverables.
Get in touch to explore how this approach can support your foresight and insight agenda.