The setup
You are about to raise the price of your main plan by roughly twenty percent. The decision is already made on the spreadsheet side — the margin needs it — but nobody in the room can say with confidence what the announcement does to the customer base that has to absorb it.
This scenario stages that moment. The cast is four audiences with different stakes: long-tenure power users who feel ownership over the product, newer subscribers who joined on price, churned users who enjoy saying “told you so,” and the forum regulars who set the tone everyone else copies. The trigger is your actual announcement — email plus changelog post — landing on simulated day one. The horizon is two weeks, which is long enough for the first wave of anger to either burn out or organize.
What makes pricing announcements worth simulating rather than guessing at is the second-order behavior: customers don’t just react to the number, they react to each other’s reactions. A multi-agent run captures that contagion in a way a survey cannot, because the agents read and answer one another round by round.
The prompt
Copy this into MiroFish as your scenario question, swapping the specifics for your own:
Our SaaS product raises its Pro plan from $29 to $35/month, announced by email and a changelog post on day 1. Simulate two weeks of reaction across four groups: long-tenure power users, price-sensitive new subscribers, churned users watching from the sidelines, and community forum regulars. Tell me which group objects first, whether the objections harden into one organized narrative, what the split between acceptance and anger looks like by day 14, and what a founder reply on day 2 would change.
Seed files that help
- Your current pricing page plus the draft announcement email — the exact wording is what the agents will react to, so use the real copy.
- A sample of recent support tickets or forum threads that mention price, so the knowledge graph learns who complains and in what tone.
- A one-page competitor price sheet, which lets the simulation treat switching as a live option instead of an abstract threat.
What to look for in the report
- First-mover segment: which group posts the first negative reaction, and how quickly the others adopt its framing.
- Narrative compression: whether scattered complaints merge into one repeatable sentence — 'they punished loyal users' travels much further than a dozen individual gripes.
- The acceptance curve: the round where anger stops growing. If it never flattens, the problem is usually the announcement copy, not the number.
- Recoverable paths: what the report says a grandfathering offer, a founder reply, or a two-week delay would have changed.