Remember the decision
Keep the agreed choice with its rationale, assumptions and review conditions—not buried in a presentation or chat.
Outside Chair — Strategy · Free to start
Know what was decided. Keep it moving.
Keep strategic decisions, reasons, assumptions and outcomes connected across every review—in the AI you already use.
Less reconstruction. Better decisions.
Walk into every strategy review knowing what matters.
Ask ordinary questions. Strategy connects the approved priorities, evidence and decisions needed to answer them.
Start where you left off
See what changed since the last approved review instead of reconstructing the strategy every time.
Know what needs attention
Bring overdue actions, challenged assumptions and stale measures into one concise view.
Keep evidence honest
Separate the source, its claim and the team's interpretation so uncertainty remains visible.
Preserve disagreement
Record attributed viewpoints and dissent without rewriting them into artificial consensus. Inviting collaborators requires Founding Pro; solo use remains available on Free.
Stay in control
Your AI can read and propose. Only an authorised person can approve a governed change.
One useful rhythm
Bring the material. Ask what changed. Approve the record.
- Bring what you haveUse existing plans, reports, board notes and decisions.
- Ask what mattersSee what changed, what is overdue and what remains untested.
- Make the decisionKeep the choice, rationale and conditions together.
- Confirm the recordNothing governed changes until an authorised person approves it.
Worked example · prompts only
Run one small strategy all the way through.
LunchLoop is a synthetic office-lunch pilot. Paste these prompts in order to see Strategy frame the question, challenge the assumption, govern execution and carry the learning into the next review.
01 FRAME
What are we trying to prove?
Prompt
Create a private project workspace called LunchLoop. Prepare—but do not save—a baseline with this priority: prove sustainable lunchtime demand for a 15-minute office-meal service; this initiative: run a 10-working-day pilot for one office cluster; this assumption: employees will pay CHF 16; this measure: paid orders per day; this signpost: at least 30 paid orders per day by day 10; and an unresolved choice between one fixed price and a split-price test. Link the records and show me exactly what will be written.
A connected baseline, not a loose summary.
- 6 versioned records
- 7 explicit relationships
- Nothing saved before your confirmation
02 CHALLENGE
What could change our mind?
Prompt
Add three prompt-only evidence cards to the CHF 16 assumption: supporting—12 of 20 interviewees regularly buy lunch above CHF 15; contradicting—nearby lunch offers cluster at CHF 13–15; contextual—delivery speed matters more on meeting-heavy days. Then prepare a pre-pilot review that preserves the tension and recommends a CHF 14/CHF 16 split-price pilot. No uploads.
Evidence stays distinct from interpretation.
- Support, contradiction and context
- Sources pinned to the assumption version
- A review approved in two deliberate steps
03 DECIDE
What will we do now?
Prompt
Prepare a decision to run a 10-working-day CHF 14/CHF 16 test, alternating price by day and tracking paid orders and repeat purchase separately. Create two actions: Ops Lead launches the menu and 15-minute logistics; Growth Lead configures the landing page and per-price tracking. Stop before confirmation. After launch, mark the Ops action complete.
The decision turns into owned follow-through.
- Immutable decision snapshot
- 2 confirmed actions with owners and deadlines
- Versioned completion with its rationale intact
04 LEARN
What did reality teach us?
Prompt
Record these synthetic pilot results: 34 paid orders per day, 21% repeat purchase and 92% on-time delivery; CHF 14 produced 46 orders but inadequate margin, while CHF 16 produced 22 orders with sustainable margin. Record a mixed outcome, challenge the CHF 16 assumption, and prepare an outcome review compared with the pre-pilot review.
The next review starts from learning, not memory.
- 2 confirmed observations and a mixed outcome
- Assumption revised from untested to challenged
- Review-to-review comparison and a CHF 15 follow-on test
Signal
34 orders/dayThe aggregate signpost was crossed.Tension
Demand ≠ marginNeither tested price solved both.Learning
CHF 16 challengedProbability revised to 0.35.Next move
Test CHF 15Segment by delivery-speed context.This worked example was run end to end against Outside Chair — Strategy 0.7.5. Its figures and sources are synthetic; no document upload is needed.
A useful first question
“What changed since our last review, what needs attention now, and what are we still guessing?”
Install version 0.7.10
Choose the AI you already use.
These are the official immutable Strategy packages. The same Outside Chair account works across supported hosts.
ChatGPT
Workspace uploadRequires an eligible workspace with plugin upload access. Ask your workspace owner if Upload plugin is unavailable.
Verify download
outside-chair-chatgpt-0.7.10.zip · 19.5 KB
4cd4bb80e71524a7f4e9489487bb8a20a335c64db26143146d756e68693de945Claude
Paid Claude planRequires a Claude plan and workspace that permit custom plugin uploads.
Verify download
outside-chair-claude-0.7.10.zip · 22.8 KB
a092a645dfa1bc73ec0fa557d7373802e0edc47ae1d0f7d9b44afb3502ecd8b0Microsoft 365 Copilot
Admin installRequires a tenant that permits custom agents and an administrator to upload and assign the package.
Verify download
outside-chair-microsoft-365-copilot-0.7.10.zip · 44.0 KB
0d5b7a7e02e22efd83103995eb3daf6e012f8a3f7e3238075863d3b6e7912614One suite price
Strategy, Performance, Task and Knowledge. $10 a month with Founding Pro.
Get up to 30 active workspaces across the suite. Join while it is available and keep that monthly price while your subscription remains active.