Outside Chair

Outside Chair — Performance · Free to start

Know what you achieved—and what the evidence really says.

Turn the KPI workbooks and notes you already use into a verified review of what changed, what needs attention and what the evidence supports.

Install Performance Try the walkthrough

Less reporting theatre. Better questions.

Turn the monthly pack into a decision-ready review.

Performance verifies the evidence, keeps uncertainty visible and never invents a cause the documents cannot support.

Start with the pack you have

Bring the spreadsheets, KPI packs and management notes already used by the team.

See only what matters

Separate decision-relevant changes from routine movement and reporting noise.

Verify every figure

Trace a displayed value to its period, comparator and source location.

Question the evidence

Ask about a fact, comparison, breakdown, claim or commitment without rebuilding the review.

Check what management says

See whether a claim is supported, contradicted or directionally right without proving its cause.

Refuse unsafe conclusions

If definitions, units or periods cannot be verified, Performance asks once or says it cannot verify the answer.

One useful rhythm

Upload the pack. See what changed. Ask the evidence.

  1. Bring the filesUse the workbook and management note the team already has.
  2. Get the conclusionReceive a concise review led by the outcome that answers your question.
  3. Inspect the evidenceCheck comparisons, components, claims and source traces.
  4. Return next monthSee only what changed and whether a commitment moved as expected—without claiming causation.

Worked example · prompts only

Run one operating review all the way through.

LunchLoop is a synthetic office-lunch company. Paste these prompts in order to see Performance establish the model, close the month, govern a response and separate the result from attribution.

  1. 01

    SET UP

    What will we manage consistently?

    Prompt

    Create a private workspace called LunchLoop Performance. Prepare—but do not save—a monthly model with three higher-is-better metrics: on-time delivery rate, repeat order rate and contribution margin per order. Use this synthetic history: August 93%, 56%, CHF 6.80; September 91%, 54%, CHF 6.30; October 88%, 50%, CHF 5.40. Use management comparators of 95%, 60% and CHF 8.00. No documents. Show the complete proposal first.

    A governed operating baseline, not another spreadsheet copy.

    • 3 defined metrics and 12 observations
    • Currency, cadence and direction preserved
    • Nothing saved before exact confirmation
  2. 02

    CLOSE

    What changed, and what needs attention?

    Prompt

    Prepare—but do not confirm—the November close: on-time delivery 90%, repeat orders 52% and contribution margin per order CHF 6.10, against comparators of 95%, 60% and CHF 8.00. Route clustering began mid-month; two large corporate events increased demand volatility. Treat that as context, not proof of cause. Show the findings and limitations first.

    The month improves, yet two material gaps still need attention.

    • Month-on-month and comparator views
    • Repeat orders and margin open material gaps
    • Short history and missing prior-year evidence stay visible
  3. 03

    COMMIT

    What response will management own?

    Prompt

    For the repeat-order gap, prepare a commitment for Maya Chen, Head of Customer Operations: call each first-time customer within 24 hours, record the non-repeat reason and test one recovery offer. Expect a 3–6 point increase by the December outcome review. Preview the evidence design first, then prepare the proposal and stop before confirmation.

    An owned response with its evaluation rules frozen before results arrive.

    • Named owner, baseline and evidence-ready date
    • Minimum meaningful effect and 90% interval
    • Low detectability disclosed instead of hidden
  4. 04

    FOLLOW UP

    Did the result improve—and can we credit the action?

    Prompt

    Prepare a December follow-up: repeat order rate reached 56%, up 4 points and inside the expected range, but below the 60% comparator. The team completed the pilot. A holiday order-mix shift and a 10% recovery coupon were material concurrent changes. Record a mixed result, keep attribution separate and complete the one-period commitment. Stop before confirmation.

    Progress is recorded without turning correlation into a success story.

    • Mixed business result: +4 points
    • Attribution remains not isolatable
    • Commitment completes; the original close stays intact

Baseline

3 governed metricsFour months stay comparable.

Close

2 material gapsImprovement does not erase the shortfall.

Result

+4 pointsRepeat orders reach 56%.

Judgement

Not isolatableConfounders prevent a causal claim.

This worked example was run end to end in a fresh ChatGPT Desktop task against Outside Chair — Performance 0.1.33. Its figures and sources are synthetic; no document upload is needed.

A useful first question

“Here is this month's pack. What did we achieve, what changed, and what decision do we need?”

Install version 0.1.33

Choose the AI you already use.

These are the official immutable Performance packages. Microsoft can use saved records; reviewing a new raw workbook requires an extraction-capable host.

ChatGPT

Workspace upload
Download ZIP

Requires an eligible workspace with plugin upload access. Ask your workspace owner if Upload plugin is unavailable.

Verify download

outside-chair-performance-chatgpt-0.1.33.zip · 27.9 KB

ff75104fd798c5c61a1f68a62fe53bb5594371628c53eab20676249baa76dacb

Claude

Paid Claude plan
Download ZIP

Requires a Claude plan and workspace that permit custom plugin uploads.

Verify download

outside-chair-performance-claude-0.1.33.zip · 26.9 KB

a8101a423b26698b7e0380e100dbf5ab856c5b56d079d4e204a5ba8df21650ec

Microsoft 365 Copilot

Admin install
Download ZIP

Existing saved Performance records, commitments and follow-ups are supported. A new raw-spreadsheet review needs a host that can run the bundled extractor.

Verify download

outside-chair-performance-microsoft-365-copilot-0.1.33.zip · 42.7 KB

db2166f64ba995b39d30b285b9e4fb7b2b9900f85d043dbed21a50c814c6afbf

One 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.

Compare plans