AI Dominates the agenda at esomar Congress 2025, Prague
Elaine Rodrigo presenting at esomar Congress
Ray Poynter, 1 October 2025
The 2025 ESOMAR Congress in Prague was dominated by one theme above all others:, and that theme was artificial intelligence. From the exhibition floor to the presentation sessions and corridor conversations, AI was the connective tissue running through the programme. What had been emerging in previous years became fully mainstream in 2025. Far from being a specialist topic, AI now underpins the methodologies, business models and strategic debates across the market research and insights industry.
This year’s Congress showcased this transformation vividly. Exhibitors emphasised AI capabilities, whether in data collection, analysis, or reporting. Presentations explored both opportunities and anxieties, ranging from efficiency gains to ethical pitfalls. The breadth of applications made clear that AI is not a tool for a niche segment, but rather a platform upon which the next phase of insights is being built.
The Exhibition: AI as Default
Walking the exhibition hall (which was enormous in Prague), I encountered AI at every turn. There were specialist AI companies, such as conversational research platform Yasna, along with a fantastic range of AI specialisms.
Even suppliers of more traditional solutions, such as the survey platform QuetionPro, have shown that they integrate AI into their workflows. This ubiquity reflects a shift: AI is no longer an optional add-on but an expectation. Clients now expect suppliers to deploy AI to drive speed, scalability, and cost-effectiveness. I heard there were 72 exhibitors and that only two were not displaying some aspect of AI.
Conference Themes: From Concept to Practice
The papers and presentations reinforced this sense of transition from theory to practice. A few key themes stood out:
1. Synthetic Data as a Practical Solution
Synthetic data received considerable attention. The debate is no longer about whether synthetic data has a role, but about where and how it can be trusted. Practical case studies demonstrated the use of synthetic data, including in the context of the global brand Reckitt. The session made clear that adoption is accelerating, but governance frameworks are needed to ensure transparency and validity.
2. Generative AI and Storytelling
Several papers explored how generative AI is reshaping reporting. Tools that summarise interviews, produce draft reports, or even generate visual presentations are already embedded in agency toolkits. One striking example demonstrated how an AI engine generated instant, “executive-ready” summaries of multi-country qualitative outputs, reducing reporting time from weeks to days.
Yet there was caution as well. Delegates raised concerns about accuracy, loss of nuance, and the risk of homogenisation. The consensus was that AI can accelerate production but must remain under human oversight, particularly when insights drive strategic decisions.
3. AI and Research Ethics
A recurrent strand was ethics. The proceedings include multiple contributions that stress the need for responsible AI, with a focus on data privacy, bias, and explainability. Discussions extended beyond technical bias into societal impact. Several speakers emphasised the responsibility of insights professionals not only to utilise AI, but also to examine its outputs and implications critically.
4. Integration Rather than Isolation
What distinguished 2025 from earlier events was the degree of integration. AI was not treated as a specialised topic area; it was woven into sessions on brand tracking, shopper research, and social analytics. For example, a retail case study combined behavioural data, AI-enhanced video analytics, and synthetic scenarios to optimise in-store layouts.
5. The Big and the Small
The three key groups that seem to be pushing AI the fastest are smaller, agile companies, big traditional companies, such as Ipsos and Kantar, and new arrivals. The main challenge might be for mid-sized companies, which lack the scale of the big players and the agility of the smaller and newer players.
Conversations in the Corridors
Beyond the formal sessions, informal exchanges confirmed the mood. Most attendees were less interested in debating whether AI should be adopted and more focused on swapping notes on implementation. For example, determining which providers to trust, integrating AI into mixed-methods projects, and balancing automation with human oversight and creativity.
One theme cutting across conversations was client demand. Buyers of research are increasingly expecting AI-driven efficiency gains (and many seem to expect price reductions), but they also want reassurance on validity and ethics.
My Session on Synthetic Data
On the final morning, I reported back on the work being done by the ESOMAR team in terms of guidelines for Synthetic Data, with big thanks to the broader team and especially to Jon Puleston and Judith Passingham, who worked with me on creating “5 Topics of Discussion to Help Buyers of Augmented Synthetic Data”. I updated the room on ESOMAR’s plans to expand the guidance to all forms of Synthetic Data, and I answered questions from the audience. The main messages were 1) Synthetic is here, 2) Transparency is key, and 3) More experimentation is required to enable buyers to make informed decisions.
You can see the presentation that Xabier Palacio and I gave on Synthetic Data thanks to the esomar live stream recording via the YouTube link below