Overview
An AI forum where a user poses a question and a panel of eight leading psychology and AI researchers each responds in character, followed by a synthesis of where they agree and diverge.
The challenge
Single-voice AI answers flatten real disagreement. The goal was to show how different expert frameworks would approach the same question, then bring the views together without losing the nuance.
The approach
- Designed eight distinct researcher personas, each grounded in that researcher’s published perspective
- Orchestrated parallel Claude API calls, followed by a synthesis pass
- Gave the interface a warm, literary treatment suited to long-form reading
The outcome
A working demonstration of multi-agent AI design, the same architecture CDG applies to client tools that need structured, explainable AI output.
Mind Hive runs on the visitor’s own Anthropic API key, so the live version is shown here by description. Contact us for a guided demo.