The CodexMR team spent four days in Valencia for ESOMAR Congress 2026. We came home with more questions than we left. It’s not a complaint. Here are three hings worth carrying home from this year’s congress.
ESOMAR Is Reshaping Its AI Guidance, Live
ESOMAR launched a new AI Alliance this year. It builds on the group’s earlier AI Taskforce. We watched one part of it, Workstream 1 on Ethics, Code and Quality, present its first guidelines live from the stage. The old approach used one checklist for everything: the 20 Questions to Help Buyers of AI-Based Services. The new approach splits that work into pieces. We saw two separate 10-question guides on screen. One is for buyers checking out suppliers. The other is for teams using AI inside their own research process. There’s also a companion guide on training data, plus a shared glossary.
That split makes sense. A team using AI inside its own process faces different risks than a client hiring an outside AI supplier. So one checklist was never going to cover both jobs well. We built our own answers to the original 20 Questions before the congress. We’ll update that work once ESOMAR’s new documents are final. For now, expect guidance that’s more specific and more split apart, not one big checklist for everyone.
The AI Conversation Has Moved Past Hype, Into Accountability
We wrote before the congress about AI fatigue among research buyers. Valencia confirmed it in person. Nobody on stage was still selling “AI-powered” as the whole pitch. The real conversations, on the floor and in sessions, were about something else: what breaks when AI gets something wrong, and who catches it first. That’s the accountability question we said mattered more before we even landed. Now it’s clearly the industry’s question too, not just ours.
Quant Research Operations Got a Real Seat at the Table
Stefka Mihaylova, our VP of Sales, spent much of the congress talking about quant research operations. That’s the programming, scripting, and QA work that has always run behind the scenes. This part of the industry doesn’t usually get its own track. But this year, it kept coming up inside the AI sessions, because it’s the part of the workflow AI touches first. Her own session, in the Emergence, Discovery track, drew a full room. Everyone was asking the same practical question: where does AI take real weight off a team, and where does a person still have to check the work?
Our answer hasn’t changed since before we landed in Spain. AI can’t do much on its own without a human on the wheel. That’s the ground CodexMR’s platform is built on.
Valencia Itself Did Half the Work
Four days is a long congress, so the sessions were only half of it. The real value still showed up in hallway conversations, at the coffee stations, over dinner, in all the unplanned moments a virtual event can’t replicate. Valencia made that easy. It’s a compact venue, the weather held up, and it’s also a city built for people to keep talking past the last session of the day. If you skipped this congress to save the travel budget, that’s the part you missed. Not the keynotes.
We’ll still be watching how ESOMAR’s new AI Alliance guidelines take shape, and we’ll update our own answers once they do. If any of this maps to a decision your team is making right now, start with our piece on how CodexMR answers the original 20 Questions. Our article on AI fatigue among research buyers covers the trust problem in more depth.
