Discovery

Discovery uses 19 structured AI audience archetypes to explore who might respond to a product and why. It generates model-based hypotheses about interest, objections, and motivations; it does not identify or measure actual customer segments.

How Discovery Works

The workflow compares model-generated responses across defined demographic, behavioral, and psychographic contexts. Outputs can include:

• stated model interest
• stated purchase-intent style responses
• objections and barriers
• motivations and questions

These outputs are synthetic. Validate important audience and purchase hypotheses with real customers or live-market outcomes.

Running a Discovery

1. Select a product from your product list.
2. Click Run Discovery or press D.
3. Configure available audience and run parameters.
4. Start the simulation.

Discovery Parameters

| Parameter | Description | | --- | --- | | Model iterations | Mathematical or model repetitions; not human sample size | | Demographic focus | Defined attributes supplied as context | | Region | Geographic context for the model run |

Understanding Results

Archetype and Segment Analysis

Discovery can group similar model responses and show:

• the share of model output in a grouping
• associated defined attributes
• model-generated affinity scores
• recurring motivations and objections

Grouping proportions are not measured population shares, and displayed segments are not discovered market segments without separate real-world evidence.

Archetype Cards

Cards show synthetic profiles and model-generated reasoning. They can help explain why the simulation produced a reaction. They are not customer records or quotations.

Competitive Reactions

Model outputs may compare a product with supplied alternatives, including perceived strengths, weaknesses, switching barriers, and questions to validate.

Best Practices

1. Supply complete, accurate product context.
2. Run sensitivity checks with plausible alternative assumptions.
3. Compare reasoning across archetypes, not just aggregate scores.
4. Export findings with model-evidence labels intact.
5. Validate promising or consequential hypotheses with real customers.

Related Articles

• Products - Setting up products for Discovery
• How AI Archetypes Work
• Compare - Comparing model runs