AI Powered Insights Conference 2026

09.00 Registration & coffee

09.30 Welcome from the Chair

09.40 AI makes ideas, but not meaning: what Diageo learned building AI innovation workflows
AI can generate endless ideas but when teams rely on the same generic models and shallow context, outputs can quickly become generic. In this session, Diageo and Zappi reveal how they moved beyond surface-level outputs by combining human skill, system design and specialised AI agents to redesign how AI fits into innovation workflows.
Hear how Diageo embedded AI into its innovation planning system, integrating it with rich connected data sources. But game-changing outputs were achieved with the addition of hypothesis-led prompting, small experimental workflows to improve results, richer connected data and agents trained on signals that drive trial and distinctiveness, resulting in faster innovation cycles and more meaningful opportunities.
Babita Earle, Managing Director, International, Zappi
Deirdre van Zyl, Head of Innovation Planning/ Insight, Diageo

10.05 Transforming research with hard-to-reach patients through AI-powered tools
Understanding patient decision-making at scale is a persistent challenge, especially when key groups are hard to reach and bespoke research is time-consuming and costly. In this session, Pfizer and CogCo demonstrate how they tackled this by building an AI-powered research platform that brings patient insight into every day decision making. They’ll share how they built interactive patient profiles based on qualitative and quantitative data which teams can use to explore perspectives, test ideas and access evidence in real time. They’ll also show how a behavioural copilot within the tool surfaces the hidden behavioural barriers and drivers affecting patients.
Rosie Foster-Carter, Principal Behavioural Scientist, CogCo
Owain Service, CEO and Co-Founder, CogCo
Edward Gardiner, Managing Director and Co-Founder, CogCo
Sarah Holiday, Director, Research and Behavioral Science, Pfizer

10.30 Designing decision-grade simulated consumers: lessons from No7 Beauty Company 
Innovation teams need to explore and refine ideas quickly, often before formal research. No7 Beauty Company sought a way to embed consumer thinking earlier. 
In partnership with Verve, they are accelerating innovation by turning segmentation into AI-powered segment experts. Trained on real concept testing data and shaped by researcher-led frameworks, these simulations reflect how audiences evaluate ideas in practice. Applied across consumer tensions and product ideation through to claims, packaging, product naming, and comms development, they are grounded in real behaviour and validated against outcomes - giving teams confidence to make smarter, faster decisions. 
This session shares practical lessons on building, validating and applying simulated consumers to drive smarter, faster innovation. 
Kylie Fitpatrick, Director, Verve   
Daniela Lipkin, Associate Director, Verve 
Rosie Hudson, Global Insight Manager, No7 Beauty Company 

10.55 Refreshments

11.25 Building an AI consumer panel from Sky data
In this case study, Sky shares its approach to creating synthetic consumer panels using qualitative focus group data and AI generated segmentation. Explore how the panel was developed, how it performs against key use cases and what it reveals about the potential of AI-generated audiences. Discover how these tools are reshaping the way consumer insights are generated, tested and brought to market.
John-William Awbrey, Head Of Brand & Campaign Insights, Sky
Marc Olm Serra, Lead Data Scientist, Sky

11.50 Weaving qual into quant: what AI interviewing really delivers
Can AI make traditional surveys feel more human and help respondents open up more candidly? In this session, Flood and Specsavers share how combining a robust quant survey with AI-led interviews of 400+ respondents unlocked deeper insight into customer trust. Discover why participants opened up more to an AI interviewer and examine best practice thinking around how to get more from quant with conversational AI methods. You’ll hear practical lessons and an honest reflection on where AI fell short.
Luke Cadman, Senior Partner, Flood
Emma Hart, Junior Partner, Flood
Iain Stanfield, Head of Insight UK&I, Specsavers

12.15 Stop using LLMs to identify your next big innovation opportunity
AI is everywhere in innovation, yet many teams still lean on large language models as idea?generation shortcuts, overlooking the risks, limits and controls needed for responsible use.
In this session, Kantar x Quilt.AI share an AI?native, agentic approach that moves AI from idea generation to insight generation. By combining cultural evidence, behavioural science and market signals, the system detects patterns and emerging opportunities at scale – before LLMs help bring early concepts to life.
Early case studies reveal common pitfalls, essential guardrails, and why blending human judgement with machine intelligence delivers stronger, more differentiated innovation.
Dr Nicki Morley, Global MD Innovation solutions, Kantar
In conversation with
Dr Angad Chowdhry, Co-Founder & CPO, Quilt.AI
Client TBC

12.40 Lunch

13.40 Panel: How much should we trust AI analysts? The role of researchers in guaranteeing quality
AI analysts are delivering faster and more accessible insight than ever before but to what extent can we trust what they deliver and what is the researchers’ role when they’re at work? This session addresses the most pressing questions facing researchers today around data quality, transparency and accountability for AI-generated outputs.
Using learnings from recent pilots and live implementations panellists will share how they validate findings, maintain methodological rigour at speed and retain oversight of increasingly agentic workflows. They’ll candidly discuss what experience has taught them about the boundaries of AI capability, when human judgement must take the lead and how the researcher’s role is evolving from analyst to manager of insight workflows.
Chair: Mike Stevens, Founder, Insight Platforms
Panel:
Matthew Picton, Head of Research, Dubit
Leon Hubert, Lead User Research Consultant, Hippo Digital
Lucia Juliano, UK Head of Research, Toluna

14.10 Beyond efficiency: what AI-augmented semiotics taught us about the irreplaceable researcher
Explore what happens when semiotics is scaled through an agentic AI system and where human expertise remains indispensable. Quantum, in partnership with Unilever, developed a multimodal AI framework to identify cultural signals, map emerging codes and generate fragrance innovation platforms at speed. While AI proved highly effective in broad discovery, rapid hypothesis generation, and cross-market analysis, it fell short in areas requiring true cultural interpretation.
This session shares the mechanics of the pilot, the honest failure modes encountered, and the operating model built in response. What do these learnings mean for how AI-augmented research is designed going forward?
Hannah Robbins, Lead, Quantum CS
Ben Lewis, Insight Lead – Homecare, Unilever

14.35 From self-serve to self-drive: how agentic AI is transforming insight democratisation across Microsoft and increasing organisational capability
Microsoft’s Business Sales Operations (BSO) team was overwhelmed by internal queries, limiting time for higher-value work. To address this, Ipsos combined AI, automation and analytics to power ‘George’ a self-service chatbot solution for expediating insight delivery to stakeholders and democratising insight across Microsoft. The solution has been adopted globally in 2025 transforming insight delivery and business models.
This talk will outline the comprehensive AI roadmap for 2026 which includes building data & AI agents to enable users to query data, interpret results, and uncover insights with natural language simplicity.
Rhys Harmour, Commercial Consultant, Ipsos Data Labs
Joseph Brooker, Senior Business Program Manager, Microsoft

15.00 Refreshments

15.25 The researchers didn't get replaced, they got a multiplier
Thirteen months ago, The Times set out to increase research quality and volume without overburdening its panel or team. The solution was not a technology replacement story but a rethink of how human judgement and synthetic intelligence work alongside each other.
Working with Electric Twin's synthetic audience platform, The Times has run the equivalent of 10x its previous annual research cutting decision times from weeks to seconds.
But the biggest shift has been cultural. This session explores how trust has been built, through validation, transparency and human oversight and what it takes to embed a hybrid human / AI research model.
Leanne Tomasevic, Insights & Strategy, Electric Twin
Chris Courtney-Smith, Director of Data Operations, News UK, The Times

15.50 Panel: MRS Delphi Group AI report
Stay ahead with the latest thinking from the MRS Delphi Group’s newly published report.
Speakers to be announced

16.20 Panel: Stop speculating. Here's what the AI-ready researcher actually looks like
This cross-industry session will introduce the new generation of AI-native researchers doing work that barely existed two years ago. Rather than running traditional research projects, their job is to build and stress-test the AI systems that generate insight. They are less analysts, more architects of the environments where insight emerges.
Explore what the AI-ready researcher looks like in practice: how they think, how they work and how quickly they are having impact. Consider what these roles suggest about the future of entry-level research and how hiring, training and team structures may evolve as AI reshapes the craft of insight.
Chair: Kelly McKnight, Executive Director, Verve
Panel:
Katie O’Connor, SVP, Behaviourally 
Liz Norman, CEO, Elizbeth Norman International 
new-entrant working in an AI-focussed role at Verve 
new-entrant working in an AI-focussed role at Behaviourally 

16.50 Closing comments from the Chair

17.00 End of conference