Draft 10 new ventures. Kill 9 by Friday.

What the AI will do

• Find problems, capabilities and market evidence, then draft ten genuinely different ventures.
• Investigate each venture’s decisive uncertainties, eliminate nine against agreed criteria and design the survivor’s first real test.

What you can provide

• A business area, capability or ambition to explore, even if it is still rough.
• Access to relevant company evidence and any limits on investment, timing or what kind of business you want to build.

Where you'll take part

• Set the decision date and agree what would make a venture worth pursuing.
• Resolve consequential tradeoffs and choose the surviving venture—or conclude that none earns a test.

What you'll get

Ten concise venture briefs, a clear record of nine eliminations, and a practical test for the survivor. If none clears the agreed bar, you get a supported no-go conclusion.

Activity context

Ten ventures are enough to loosen your attachment to the first good idea. The deadline gives the exercise teeth: by Friday, nine leave the room. The survivor earns a small real-world test, not a business plan built on enthusiasm.

For the user

Start with the territory you want to explore and whatever constraints matter. The AI will look for evidence, draft ten distinct ventures and investigate the questions that could kill each one. You will agree the decision criteria and settle the final choice; none of the ten has to survive if the evidence is weak.

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

Help me draft ten new ventures, eliminate nine by an agreed decision date and prepare the survivor's smallest useful real-world test. If none deserves that test, reach an explicit no-go conclusion rather than manufacturing a winner.

1. Establish the territory and decision

Read what I supplied and establish the business area, ambition and decision date; use Friday if that is the intended deadline, clarifying which date where necessary. Inspect and try access to strategy documents, customer interviews and support records, lost-deal notes, abandoned proposals, innovation reviews, financial and operating data, and relevant email, chats, WhatsApp exchanges and meeting transcripts. Search by the customer problems and capabilities in my starting material. Look for recurring unmet needs, work customers already pay to do, valuable capabilities we possess, unused capacity, previous experiments and why earlier proposals stopped. Use records and dates to distinguish an idea we rejected from one we simply never investigated.

Identify potential buyers and users from account records, meeting participants, project roles and their own correspondence. Record their expressed problems, current alternatives, constraints and purchase process; keep inferred motivations separate. If source access is missing, inspect connection and export options and help me supply a few permitted examples. Use a narrower hypothesis-led exercise if needed, carrying that weaker evidence into elimination decisions.

Ask focused questions one at a time about the consequential gaps: what sort of business we want, resources we can commit, necessary strategic fit and conditions that would make a venture unacceptable. Propose a small set of criteria and agree them before eliminating ideas. Distinguish hard requirements from preferences; do not invent numerical precision where judgment is the honest measure.

2. Draft ten genuinely different ventures

Use the evidence to explore materially different customers, problems, offers, delivery models or ways of earning revenue. Ten names, sectors or feature variants on one product do not count as ten ventures. Write a short brief for each with:

• A specific customer and problem, including what the customer does today.
• The concrete thing we would sell and why someone might switch or pay.
• How it would reach customers, deliver the result and earn money.
• The relevant capability or advantage we could bring, with its evidence.
• The hardest assumption and the smallest observation that could disprove it.

Where useful, give separate agents bounded discovery briefs with the same evidence and criteria but different genuine search territories. Have them propose ideas independently before seeing each other's concepts. Reconcile overlaps and fill missing territory until there are ten distinct briefs. Retain source references and label creative proposals as proposals. If agents are unavailable, use bounded passes and disclose shared context. Show me the ten together and incorporate corrections to facts or scope before investigation.

3. Investigate what could kill each one

For every venture, choose the one or two uncertainties most likely to change the decision: an already adequate alternative, no reachable buyer, implausible cost to deliver, absent distribution, a missing capability or a dependency we cannot obtain. Write the question and the evidence required to answer it. Research current alternatives, customer behavior and purchasing, delivery requirements and cost drivers using relevant original sources such as company product and pricing pages, filings, procurement documents, published customer research and our actual records. Do not substitute a large market-size number for evidence that this customer has this problem.

Divide the ten investigations into bounded assignments. Give each agent the relevant venture brief, source packet, exact decisive questions, agreed criteria, time available and required output: supported findings, contrary evidence, unresolved questions and a continue/kill recommendation with reasons. Keep a coverage record of all ten and use successive waves if capacity is limited. If source material or task size exceeds a session, narrow the question or split the sources. The lead checks important original evidence and reconciles disagreement after each independent investigation finishes.

Use ranges for uncertain costs and revenue assumptions, making the units and missing inputs clear. Do not fabricate customer interviews, current competitor terms or a test we have not run. Where an assumption cannot be resolved by desk research before the deadline, identify the real observation needed and judge whether the remaining uncertainty merits a test.

4. Eliminate nine

Compare all ten against the agreed criteria and the actual findings. Start with hard failures, then compare the strongest remaining alternatives using a common account of customer value, access to the buyer, delivery feasibility, economics and our reason to win. Explain tradeoffs in plain language rather than hiding them inside a weighted total.

Give a fresh reviewer the ten briefs, criteria and evidence without the authors' preferred winner. Ask it to challenge unsupported elimination reasons and the strongest surviving claim. Reconcile substantive findings; use a disclosed separate pass if fresh review is unavailable.

Present the proposed survivor and a concise, specific reason for killing each of the other nine. Distinguish evidence against an idea from the decision that another deserves our limited attention. Let me correct important facts and resolve a final consequential tradeoff. By the decision date, make the selection explicit. Do not keep nine ventures alive as “phase two” to avoid making the choice. If none clears the minimum bar, recommend no-go and say what evidence, if any, could reopen that conclusion.

5. Give the survivor a real test

For the selected venture, identify its most consequential remaining assumption and design the smallest test capable of changing our mind. Specify the actual target participant, offer or prototype, representative task or commitment, observable outcome, success and failure criteria agreed in advance, duration and limited resources required. A complimentary conversation alone may test comprehension without testing purchase; match the behavior to the hypothesis.

Create the materials the test needs within available capabilities: for example a working demonstration, neutral interview guide, draft landing page, sample service deliverable or costed pilot offer. Test the materials' basic function and preserve uncertain promises as proposals for my review. Prepare the exact next action and owner. Obtain authorization before public launch, outreach, spending or commitments; this activity's completed result is the prepared test, not invented customer results.

Deliver the ten briefs, source-linked findings, nine elimination decisions and the survivor's ready-to-run test. For a no-go, deliver the ten decisions and the evidence behind the conclusion without creating an unnecessary test. Finish at the agreed decision, with no automatic venture launch or continuing incubation of the discarded nine.

Done when

You have ten distinct venture briefs and an explicit decision eliminating nine, with a prepared real-world test for the survivor. If none meets the agreed threshold, you have a supported decision to pursue none of them.

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.