Three ingredients
Our expertise: twenty years of applied creativity
Rather than a generic prompt, CPS (Creative Problem Solving) and Design Thinking are encoded into agents, together with our professional expertise: each one applies a precise method, at a precise step.
A co-creation principle
The clarification phase enriches the context with you and sharpens your creative challenge. At every step, you choose what the AI does on its own, and what you want to add or select.
The creative agents then draw on this context to generate ideas. That's RAG (retrieval-augmented generation): answering by first searching a rich, structured knowledge base.
Under the hood, orchestration at every step
The agents don't work haphazardly: a method tells them what to do, in what order, and checks the result. That is what guarantees a relevant outcome that stays on your topic. The methods in detail
What research says
of the ideas produced by teams of AI agents are judged more creative than human ideas.
The gap comes from novelty: on usefulness, the two are level. In the top 10 of every challenge, 100% of the ideas came from the agents.
Six creative challenges, a single idea per team, limited time on both sides: the comparison is not about volume.
Hu et al., “Multi-agent AI systems outperform human teams in creativity”, Microsoft Research, arXiv:2605.17885 (May 2026). Blind rating by 5 independent judges, effect size d = 1.50.
The LaPEA (Université Paris Cité, Todd Lubart) and the IA . Innovation Club are currently evaluating the performance of AI Innovation Box. First finding: already, the generated concepts are more structured, clearer and richer than concepts generated by humans alone. Work continues on the originality and usefulness criteria.
ia-innovation.club