Conference Agenda

Synthetic Users – AI-generated personas that simulate user feedback – are gaining rapid traction in UX practice. But what are we actually measuring when we test with users who never existed?

This talk takes a critical but pragmatic look at Synthetic Users. It examines why LLM-based simulations structurally tend toward sycophancy and statistical averages, why this makes them unreliable for validation and innovation – and where they can, under the right conditions, add genuine value.

The presentation covers: the current tool landscape and what vendors promise; the sycophancy trap and why 98% of synthetic users answer "somewhat satisfied"; the fundamental difference between AI interpolation and real human behavior; legitimate use cases (Chattable Personas built on solid qualitative foundations, early-stage concept testing); and the implications for UX research practice – redirecting human research toward exploration and summative evaluation, where real people are irreplaceable.

Current conversations in UX Research are often centered around how researchers use AI tools through the research process. But what about the other human in our research studies: the research participant?

AI-savy research participants are already leveraging AI-assistants, agents and AI-tools to join, support or independently complete remote research studies - such as sending a voice assistant to an interview. So is remote research actually still researching human experiences? Is the reliability of remote insights decreasing?

As actionable take-away the presentation will cover examples on how participants use AI-tools during research studies and provide tips for researchers to identify or prevent it.

It is also an invite for reflection on limitations and risks of remote research methodologies. Hence, the content is relevant for anyone conducting remote research, especially if using participant recruitment providers.

Komplexe Produkte, Zeitdruck im Service und fragmentiertes Wissen stellen Industrieunternehmen vor große Herausforderungen. In diesem gemeinsamen Vortrag zeigen Bosch Rexroth und UID, wie der Hydraulic Hub durch konsequente User Research den Service neu denkt: weg von informationsgetriebener Suche, hin zu kontextbezogener, prozessintegrierter Unterstützung inkl. KI-gestützter Assistenz. Anhand konkreter Einblicke aus Praxis diskutieren wir, wie nutzerzentriertes Design Effizienz und Akzeptanz im industriellen Umfeld nachhaltig verbessert und die Handlungssicherheit auch für die Zukunft sicherstellt.