Case study · Product Designer · Simulation, Automation.
Thinkwell automates congressional advocacy campaigns and uses deep
persona and persuasion models to simulate how different audiences will
receive and react to a message before it ever reaches them.
The Problem
Policy advocacy runs on judgment. Teams spend weeks researching
issues, drafting materials, and tailoring messages to different
audiences with no way to know if any of it will work until it is
already in front of a congressional office or in the press.
Entry Point
Early users were immediately met with an automation grid, empty
fields, and no direction. The fix was to reduce the entire entry point
to a single question.
The Canvas
Every project resolves into the same structure: inputs at the top, an
operation in the middle, outputs at the bottom.
Conversation
Every block in the canvas has a chat panel. Instead of forms or
dropdowns, configuration happens conversationally.
Simulation
Each persona is sampled, not invented. Demographics, psychographics,
cognitive state, stress levels, persuasion filters, all drawn from
probability distributions grounded in real population data.
Conviction
We believe the gap between writing something and understanding
whether it works can be closed entirely.