Simulate an audience

Launch should not be the first test. Build a persona from analyzed behavior and language, instantiate a simulated person, and ask how it would react to your draft before a real person ever sees it. A wind tunnel for messages.

Model, instantiate, simulate

Three steps, all in windfall.audiences: build a persona from content you have already analyzed, instantiate a person from it, then simulate that person’s response to your draft.

# Build a persona from analyzed content (e.g. a thread you scored)
persona = client.windfall.audiences.create_persona(
    from_content=thread.id
)

# Instantiate a simulated person from that persona
person = client.windfall.audiences.create_person(persona_id=persona.id)

# Ask how they would react to your draft
reaction = client.windfall.audiences.create_person_simulation(
    person_id=person.id,
    content="Draft of your message here...",
)

print(reaction.predicted_response, reaction.velocity_deltas)

predicted_response is what the modeled reader would say back; velocity_deltas is how your draft shifted each of the eight persuasion dimensions for them. Test the draft against a modeled skeptic, a modeled ally, and a modeled undecided before you send it.

You have seen the whole loop: listen, understand, simulate, deliver. Head back to Getting started, or the documentation home.