Test your strategy against different scenarios.

What the AI will do

• Recover the strategy, decision horizon and assumptions, then research the uncertainties that could change its performance.
• Develop a small set of coherent possible futures and test the same strategy against each.
• Compare potential adaptations and identify observations that would tell you when to reconsider a choice.

What you can provide

• A strategy, decision or rough description of one, with access to relevant plans, financial assumptions and operating evidence where available.
• An optional accompanying scenario-planning research note; its text can be pasted if the AI cannot open it.

Where you'll take part

• Confirm the strategy and what success means, and correct the consequential assumptions before detailed testing.
• Review the scenario set and decide which adaptations or conditions deserve further attention.

What you'll get

You receive a practical comparison of your strategy across a few possible futures, with tested adaptations and a watchlist tied to actual decisions.

Activity context

A strategy can look convincing because everyone is imagining roughly the same future. Try it in a few futures that work differently: a competitor changes the market, input costs jump while selling prices fall, or an internal programme arrives too late. The exercise reveals what your strategy depends on and what you might do now to keep useful choices open.

For the user

Bring the strategy or decision you want to examine, even if it is still rough. The AI will investigate its assumptions, propose a manageable set of different futures and ask you to check them. You will then review where the unchanged strategy works, where it needs adjustment and what you would watch for.

Guidance for the AI

These are general instructions, not a description of this user's particular AI system, tools, access, or circumstances. Adapt the method to the user's goal and the capabilities you can actually use, while preserving the activity's essential sequence, review points, and intended result.

Start with the activity and the user's context

Read the complete activity and guide the user through it. Use the activity prompt yourself; the user does not need to paste it again. Keep the responsibilities of the user, the AI, and any colleague clear.

Begin with what the user has already supplied. Establish the intended result and resolve missing context through focused questions when needed. Work within the activity's stated scope, permissions, required confirmations, and stopping points, including any deliberate order for introducing information. Make progress with the available information and reasonable, clearly identified assumptions where the activity allows them.

Check the capabilities needed for this work

Inspect relevant available tools, connectors, and data. Use small, relevant, non-destructive checks to establish whether necessary access or a capability actually works. Naming a product does not establish access to it. Let this activity determine what needs checking; do not impose a universal preliminary questionnaire or a full system audit.

When useful access is missing:

• Explain what you cannot reach and how that material would help.
• Investigate practical setup, connection, export, or other access alternatives before asking the user to gather material or troubleshoot.
• If progress depends on access, request only the specific user action or permission needed to enable it or provide the material another way. Continue useful independent work while awaiting the response.
• Offer to walk the user through an export or upload. Welcome files, links, copied text, rough notes, or an interview as supported by your system and the activity, and use the best available route. If the missing access is optional, continue with what is available.

Ask focused questions and respect the answers

Ask questions only when necessary to complete the work well or satisfy the activity's required review points. Build on answers already supplied. Ask one question at a time, or a small group of closely related questions. Wait on work that depends on the answer, and continue useful work that does not. Explain unfamiliar language with examples where helpful.

Confirm answers or decisions you supply on the user's behalf before treating them as the user's final answers or decisions. This does not require separate approval of every sourced finding; follow the activity's specific review requirements.

Keep evidence and independent contributions distinct

Distinguish evidence, inference, and missing information. Check your work against the source material.

Use independent agents when your system actually supports delegation and independent investigation or review would improve the work:

• Give each agent a bounded brief and the context appropriate to its role.
• Keep discovery independent until synthesis where the activity requires it. Preserve any requirement for the user or colleagues to contribute before seeing other explanations.
• Follow the activity's stated fallback and any required confirmation when independent agents or fresh contexts are unavailable. Preserve reviewer isolation where required; describe shared-context or sequential passes accurately.

Produce, check, and revise the result

Do the substantive investigation and production. Carry the activity through to the most complete, usable result within its scope, rather than ending with a plan or handing the main work back to the user. When missing information does not prevent useful work, proceed with clearly identified assumptions where the activity permits them.

Check the result against the activity's completion criteria. At each substantive delivery, state what is complete and any material unfinished work, assumptions, helpful missing information, and remaining uncertainty. Explain how these affect what the user can use, and help the user revise the result.

Preserve the activity's stopping point. Do not turn a rehearsal, investigation, or test-design activity into execution that the user has not authorized.

Your AI prompt

1. Establish the strategy to be tested

Help me test a strategy against a small set of meaningfully different possible futures. Use an accompanying scenario-planning research note if one is supplied; try to read it and ask for its text if retrieval fails. The method below is sufficient to proceed without the companion. Scenarios describe possible conditions for learning and comparison; do not present them as forecasts or assign invented probabilities.

Inspect and try available company documents, email, chat and other messages, calendar-linked leadership meetings and transcripts. Use the strategy's business units, initiatives, decision owners and dates to find its origin, current version, revised commitments and contested assumptions. Examine relevant budgets, financial models, operating plans, programme milestones, procurement records, customer feedback and commercial data. Explain what the sources establish and what you could access.

If access is missing, inspect connection or export possibilities before asking for a specific document or action. Offer to draft a targeted access request if useful. Work with a pasted strategy, rough notes and an interview when necessary, clearly identifying assumptions that lack company evidence. Do not make access setup the entire result when a bounded strategy exercise is feasible.

Present a short testing brief: the actual decision, time horizon, intended outcomes, success criteria, current commitments, strategy components and important assumptions. Distinguish what the decision owner controls, can influence or must accommodate. A corporate programme outside a business unit's control can be a scenario condition; the unit's own proposed response is a strategic choice. Ask me to correct or confirm the brief before testing, one consequential question at a time.

2. Investigate what could change the picture

Search current primary external sources and relevant internal evidence for uncertainties with a material connection to this strategy. Depending on the decision, examine competitors' filings and capacity announcements, supplier or procurement evidence, input costs, achievable selling prices, customer behavior, access to capabilities, technology and relevant rules. Include internal execution, staffing or governance when those could determine what the strategy delivers.

For each serious uncertainty, record the present evidence with source dates, plausible divergent outcomes, the mechanism by which it affects the strategy, and what is unknown. Separate established conditions, inferred relationships and hypothetical future events. Follow interactions: cheaper materials do not necessarily help if selling prices fall faster; strong demand may not help when delivery capability is missing.

Keep the search bounded by the decision. Use agents for distinct evidence areas when breadth justifies it, with explicit questions, source requirements and completion criteria. Reconcile findings before choosing the scenarios. Use manageable passes when agents are unavailable and preserve a coverage record.

3. Construct a few coherent futures

Propose normally three or four scenarios, with a reason if a different number adds useful coverage. Choose conditions that expose different consequential assumptions, rather than optimistic, middle and pessimistic versions of the same spreadsheet. A two-axis matrix is useful when two sufficiently independent uncertainties do the work; it is optional. You may instead combine several uncertainties after checking which outcomes can coherently coexist.

For each scenario, produce a compact card with a name, horizon, short causal story, major conditions, business consequences, important interactions and the source basis for its starting assumptions. Future events remain explicitly hypothetical. Check for incompatible conditions and explain any surprising combination. Use sourced numerical ranges or clearly identified assumptions I accept; otherwise state directional effects and what remains unquantified.

For example, an invented manufacturer might face scarce materials but customers willing to pay for reliability; new rival capacity that reduces selling prices faster than costs; or strong demand while an internal technology programme stalls. These are illustrations of different causal structures, not prescribed scenarios or findings about my company.

Explain what each future adds and remove duplicates. Show the proposed set and the consequential assumptions for my correction. Wait for agreement on the set before detailed testing; keep unresolved assumptions visible if I choose to explore them provisionally.

4. Hold the strategy still and test it

Use the same agreed strategy, objectives and comparison criteria in every future. Respect each scenario's conditions rather than changing them to rescue the strategy. For every major strategy component, trace how it would work under those conditions, which assumption survives or breaks, the resulting effect on objectives and any critical dependency.

When agents are available, give each scenario analyst the same self-contained strategy and evidence packet, plus its scenario card and a common output format. Require consequences, source references, explicit assumptions, unknowns and possible adaptations. Keep first analyses independent until complete. With limited capacity, use successive waves; without agents, make separate passes and disclose shared context.

Build a strategy-by-scenario comparison. Use descriptions such as works, needs modification or fails under these conditions, with an explanation in every cell. Where transparent calculations are possible, show inputs, units, ranges and formulas; check them against the source data. Do not manufacture precision to fill a table. Keep results for the unchanged strategy separate from suggested improvements.

5. Test responses and identify signposts

Develop a few practical responses to the vulnerabilities found. Distinguish moves potentially useful across the set from conditional moves that depend on particular developments. Consider meaningful costs and tradeoffs: staging an investment may preserve flexibility but raise unit costs; building independently of a corporate programme may require additional capability.

Retest each proposed response against the same futures and criteria. Remove claims of robustness that fail this comparison. Identify what the organization would need to learn, build or arrange to keep a conditional response available, and any date after which the option becomes difficult to use.

For the consequential assumptions, propose observable signposts: the indicator, actual or potential data source, sensible review cadence, decision it could change and a threshold or event where one is supported. Do not invent company acceptance tests, dates or numerical triggers. Ask me to resolve consequential missing thresholds or leave them explicitly proposed. For example, a missed programme acceptance milestone before an equipment commitment date can trigger reconsideration; 'watch technology' is insufficient.

6. Deliver a usable strategic comparison

Check that every strategy component was considered in every scenario and that proposed adaptations were actually retested. Have an independent reviewer inspect the cards, comparison and source packet for contradictions, rescued assumptions and unsupported numbers where possible. Otherwise use a separate review pass and explain its limits. Repair the findings.

Deliver the agreed testing brief, scenario cards, unchanged-strategy comparison, the assumptions that matter most, retested adaptations and short signpost watchlist. Make the principal implication easy to find. Discuss which choices I want to pursue, distinguishing my decisions from your recommendations and any unresolved evidence gaps. Stop at the strategy review; implementation and outreach require a separate request.

Done when

You have tested the same strategy against a small agreed set of coherent futures, compared useful and conditional adaptations, and linked observable developments to decisions you may need to revisit.

Research & resources

Read the supporting research. Download research as Markdown.

These instructions were written in September 2026 using the frontier models available at the time. You may now have better tools or models, or restrictions that make some tasks harder. Your AI should adapt to the tools and access you actually have and find the best available way to help. Everything here is a starting suggestion for you and your AI, not a proven best practice for every setup. Feel free to rewrite any of it to better meet your needs.