Back to Briefs

The Synthetic Customer

A draft Signals Brief on AI-generated audiences, simulated research panels, and the strategic risk of testing decisions against synthetic demand.

Jun 12, 2026·Technology Scan·Synthetic research methods

Why this matters

Customer research is starting to split into two tracks: direct evidence from real people, and synthetic evidence generated from models, behavioral data, personas, transcripts, and market assumptions.

For innovation teams, the interesting question is not whether synthetic customers are accurate enough to replace research. The sharper question is where they begin to change the speed, cost, and politics of early-stage product decisions.

What Signals would scan

  • AI persona platforms and synthetic research panels.
  • Product teams using simulated audiences for concept testing.
  • Market-research firms repositioning around model-assisted insight.
  • Failure modes: hallucinated demand, feedback loops, overfitted personas.
  • Governance patterns for separating evidence, assumption, and simulation.

The strategic read

Synthetic customers are useful as a pressure-test layer, not as proof. The signal to watch is when teams start treating simulated feedback as a normal step in product discovery, brand testing, policy design, and venture diligence.