Skip to content
Mariete

Roll out without backlash.

Test three rollout plans. Pick the one that does not blow up.

Describe the change you want to make. Simulate runs your users through each rollout option and predicts how they will react.

6 min read
·
Intermediate
·
4 moves
·
Updated weekly
01The situation

Why this exists.

Describe the change you want to make. Simulate runs your users through each rollout option and predicts how they will react.

Today

A patchwork that breaks.

  • A shared spreadsheet nobody opens on time
  • Prompts copy-pasted into a chat window
  • A contractor who disappears for two weeks
  • Output that lands in a different shape every run
With Simulate

A workflow that ships.

  • One brief, one cadence, one place to read it
  • Every claim cited, every step reviewable
  • A finished artifact in your team’s format
  • Controversial changes roll out without bleeding users.
02The playbook

Four moves.

Simulate runs each move with a preview attached — so you know what lands before you ever hit send. Skip freely once you know which parts carry the weight.

  1. 01Kickoff

    Draft three rollout options.

    Simulate starts with the brief and asks only for what's missing. No boilerplate intake form, no setup meeting.

    Chat · mariete
    Y
    Draft three rollout options.
    S
    Simulate is working
  2. 02Gather

    Describe your users.

    Sources are pulled, cleaned, and cross-checked against prior runs — every claim carries a citation you can trace.

    checklist · step-02
    Source connected · describe your users.
    Context loaded
    First pass complete
    Source connected
  3. 03Reason

    Simulate plays out each one.

    The agent thinks out loud where it matters — trade-offs named, assumptions surfaced, judgments explained.

    ranked results
    01Roll · feature rollout0.94
    02Out · feature rollout0.82
    03Backlash · feature rollout0.67
  4. 04Draft

    Pick the option that lands softly.

    A first draft lands in the format your team already uses. You edit the last 10%, not the first 90%.

    delivery · inbox
    S
    Simulate → your team
    just now · scheduled weekly
    new
    Roll out without backlash
    Controversial changes roll out without bleeding users.
    Open briefing
03The setup

Configure Simulate.

Simulate runs on structured setup, not freeform prompts. Fill the fields once and the run is reproducible every time — same agents, same sources, same output shape.

simulate · run setup
ready
Source material
4 memos + audience brief
Agents
320 cognitive agents with belief systems
Environment
Simulated Reddit / Twitter
Horizon
7 days · 5 rounds / agent
Output
Opinion trajectories · causal traces · IMRaD PDF
Any field you skip, Simulate asks for once on first run.
Run Simulate
04What you need

Inputs in, outputs out.

Simulate runs on the inputs on the left and hands back the artifacts on the right. Skip any input — the agent will ask for it the first time it needs it.

What it takes in
  • One source of truth (CSV, CRM, or warehouse)
  • A one-paragraph brief on the goal
  • The KPI you want to move
What it hands back
  • A scored, cited brief you can forward
  • A structured file for downstream automation
  • An alert when anything material changes
05What lands on your team

A finished artifact, not a todo list.

Every run ends the same way — a packaged brief in the channel your team already reads. Here's a preview of what shows up.

playbooks·Mariete Bot·Mon 7:03 AM
delivered
S
Simulatebot · weekly
Roll out without backlash — ready for review

Here's the brief for this week. I ran the playbook end-to-end, flagged anything that shifted against last run, and packaged the output for Slack and the shared drive.

  • Draft three rollout options.
  • Describe your users.
  • Simulate plays out each one.
  • Pick the option that lands softly.
Open briefingbriefing.pdf · sheet.csv · slides.key
What moved
Best of three
Controversial changes roll out without bleeding users.
06Common pitfalls

Where teams stall.

Three ways we see this go sideways — and how to avoid each one.

Pitfall 01

Pointing the agent at stale or half-connected data. Clean the source once, compound every run after.

Pitfall 02

Running it once and forgetting. Put it on a weekly cadence so the numbers actually move.

Pitfall 03

Skipping the first review. Check the first run by hand — trust compounds from there.

07Questions

Before you start.

Usually one source is enough to see value. Simulate can run on a CSV paste for the first pass; connect the CRM, the data warehouse, or the tool of record once you want it to run on its own.

Most teams put this on a weekly cadence. That's the sweet spot between "too noisy to read" and "too stale to act on". Adjust once you see how the numbers behave.

Whoever owns the downstream action. Simulate hands back a finished result — the value is in somebody actually reading it and shipping the decision the same day.

It usually isn't. The first pass is calibration — tell Simulate what was off, rerun, and the second is close. By the fourth it reads like a teammate.

Your move

Run it today.

Forty minutes to set up. Best of three on the other side. Simulate does the work.