Cognel and research.
Preserving the capacity to reason.
We built Cognel on a conviction: the ability to reason well under uncertainty is worth protecting, and AI should strengthen it rather than stand in for it.
Whether it does is an empirical question, and the territory is new. We support the researchers investigating it, and we learn from what they find.
Understanding what actually happens.
Deciding is among the most distinctly human things people do. As AI becomes present in how choices get thought through, what actually happens to that capacity matters more than the productivity claims made about it.
The open questions are answerable. Does AI-assisted structuring improve the quality of a decision, or only the speed of reaching one? Does it surface assumptions a person would otherwise miss? Does it raise confidence, and is that warranted? Cognel is a useful instrument for asking, because every step of the reasoning is recorded.
We enter these partnerships to learn as much as to contribute. Knowing where the product helps, where it does not, and what it changes about how people work is directly useful to us.
What we offer qualifying studies.
Open to universities and academic research institutions studying AI-assisted reasoning and decision-making.
- Free licences for participants. Full access for the duration of the study, so tool access adds no licence cost or friction.
- Study design support. We can advise on using Cognel in a study context and contribute to the design where that helps. We are not a research institution, but we know the instrument.
- Custom Guides where feasible. If a study needs a decision workflow no existing Guide covers, we can build one. Subject to scope and capacity.
A Guide built for a study.
The custom Guide offer is not hypothetical. Working with Dr. Lucas Memmert, we built one to fit a study rather than the other way round. The decision task, the number of steps and the wording of each question were set by what the study needed to observe, and the reasoning it produced was comparable across participants because every one of them met the same steps in the same order.
If a study calls for a workflow that does not exist yet, reach out to us.
Working with the University of Zurich.
Cognel is built on a cognitive framework that describes productive interaction between a user and an AI, so that the result is one the user actually owns. The fundamentals of the framework are well researched. The combination is new. What it makes possible is what we want to understand. Several studies have already been carried out with two groups at the University of Zurich.
Schwabe and his team investigated design and architecture questions in the context of Cognel. The results of their studies influenced its development.
Their work looks at decision quality and how tool support and scaffolding shape it. A related paper is under peer review, and we are designing a larger study together.
Tell us what you want to study.
Tell us what you are studying, and we will walk you through the framework, how Cognel works, what it records, and where it could fit your design.
Start a conversation
Jan Fülscher
Catalina Ortiz de la Peña Svec