Speak your constraints, examples, and context into any AI tool. Wispr Flow formats it into clean, structured input — the kind that gets useful answers on the first try. No re-prompting. Start flowing free.
Eight prompts for compressing strategic thinking into days without losing rigour.
Strategy work is not constrained by time at the inputs stage. It is constrained by the discipline to structure, challenge, and choose between options. This pack compresses the structuring. Human judgement still does the choosing.
The eight prompts run in sequence. You can compress the full sprint into one focused day with two or three people, or stretch it across five days with a larger team. The sequence is the same.
What this pack assumes you know
A strategic option is a coherent direction the business could pursue, not a task. Criteria are the things you are deciding on, ranked in priority order. Stop conditions are the specific signals that would tell you to reverse or pivot. If those three terms feel unfamiliar, read prompt 5 first. It is the load-bearing one.
Suggested 5-Day Sequence
| Day 1 — Frame | market-size and competitive-map |
| Day 2 — Understand | customer-synthesis |
| Day 3 — Diverge | option-generator |
| Day 4 — Converge and communicate | strategy-choice and narrative-crafter |
| Day 5 — Commit | decision-log and execution-plan |
The Eight Prompts
market-size— The Market Sizing Promptcompetitive-map— The Competitive Mapcustomer-synthesis— The Customer Insight Synthesisoption-generator— The Option Generatorstrategy-choice— The Strategy Choicenarrative-crafter— The Narrative Crafterdecision-log— The Decision Logexecution-plan— The 90-Day Execution Plan
The Market Sizing Prompt
Purpose: Produces a defensible top-down and bottom-up market size estimate with explicit assumptions and a confidence range.
Copy and Paste
I am sizing the following market: [DESCRIBE THE MARKET IN ONE SENTENCE: WHAT IS BOUGHT, BY WHOM, WHERE] Geography: [GLOBAL / UK / EU / NORTH AMERICA / OTHER] Time horizon: [CURRENT YEAR / 3-YEAR / 5-YEAR] Anchor data I have already gathered: [PASTE ANY ANCHOR DATA YOU HAVE: INDUSTRY REPORTS, GOVERNMENT STATISTICS, COMPARABLE COMPANY REVENUES, REGULATOR HEADCOUNTS. IF YOU HAVE NONE, WRITE "NONE" AND PROCEED.] Produce a market sizing model in this exact format: 1. TOP-DOWN ESTIMATE: starting from a recognised larger market, narrow by segment, geography, and addressable share. Show each narrowing step with the assumption and the percentage applied. 2. BOTTOM-UP ESTIMATE: starting from buyer count multiplied by average spend. Show buyer count source, average spend source, and purchase frequency. 3. RECONCILIATION: where the two estimates diverge, name the divergence and the most likely reason. 4. ASSUMPTIONS LEDGER: every assumption used in either estimate, with a confidence rating of High, Medium, or Low. 5. CONFIDENCE RANGE: state the estimate as a low, central, and high figure. Explain what would have to be true for each. 6. WHAT WOULD CHANGE THIS ESTIMATE: the three pieces of evidence that, if obtained, would most narrow the range. Output format: numbered sections 1 to 6 in the order above. Constraints: - If I provided no anchor data, do not invent figures. State that the estimate is illustrative only and requires anchor data to be defensible. - Do not give a single number without a range. - Do not bury assumptions inside arithmetic. Every percentage applied must be named. - If the market cannot be responsibly sized with the information provided, say so and list what is missing.
Usage Notes
Use on Day 1. The anchor-data block is the difference between a defensible estimate and a fabricated one. Even a single industry-report figure is enough to ground the model. If you have nothing, the prompt will return an illustrative structure rather than fake numbers, which is the right behaviour.
Worked example shape
For a UK private dementia care home market sizing, the model narrows the UK social care market by private share and dementia specialism for the top-down, multiplies dementia-resident count by weekly fee for the bottom-up, flags fee-variance by region in the reconciliation, and expresses the estimate as low, central, high.
The Competitive Map
Purpose: Identifies competitors across direct, adjacent, and substitute categories. Surfaces the competitors that get missed.
Copy and Paste
My business does the following: [DESCRIBE THE BUSINESS IN TWO SENTENCES: WHAT IT SELLS, TO WHOM, AT WHAT PRICE BAND] The customer outcome we deliver is: [ONE SENTENCE: THE JOB THE CUSTOMER HIRES US TO DO] Produce a competitive map in this exact format: 1. DIRECT COMPETITORS (5 to 8): businesses selling the same product to the same customer. For each, give one-line positioning and the price band. 2. ADJACENT COMPETITORS (5 to 8): businesses selling a different product to the same customer, competing for the same budget or attention. 3. SUBSTITUTE COMPETITORS (5 to 8): different solutions to the same customer outcome, including DIY, do-nothing, and non-obvious substitutes. 4. EMERGING COMPETITORS (3 to 5): businesses not yet competing but positioned to do so within 24 months. State the path they would take. 5. THE COMPETITORS MOST FOUNDERS MISS (3 to 5): substitutes and adjacents commonly overlooked because they do not appear in the same category listings. Explain why each is missed. 6. COMPETITIVE WHITE SPACE: the customer outcomes that no competitor in the map currently serves well. Output format: numbered sections 1 to 6 in the order above. Constraints: - Do not list only direct competitors. The value of this exercise is in the adjacents and substitutes. - Do not name companies you are not confident exist. If unsure, say "I am not confident this company exists or competes as described." - Do not pad lists to hit the count. Fewer real competitors is better than more invented ones. - For each named competitor, include one verifiable detail (founding year, headquarters, known product).
Usage Notes
Use on Day 1 alongside market-size. The "competitors most founders miss" section is the load-bearing one. If the model returns only direct competitors, push back and re-run.
Worked example shape
For a UK probate-tech startup: direct competitors are other probate platforms, adjacents are will-writing services and legal-tech generally, substitutes are high-street solicitors and DIY probate. The "missed" category often surfaces banks offering bereavement services.
The Customer Insight Synthesis
Purpose: Aggregates customer interview notes into themes, contradictions, and surprising findings. Avoids cherry-picking.
Copy and Paste
I have conducted customer interviews. The raw notes are below. [PASTE INTERVIEW NOTES. INCLUDE INTERVIEWEE ROLE OR SEGMENT WHERE POSSIBLE. IF YOU HAVE MORE THAN 10 INTERVIEWS, BATCH INTO GROUPS OF 5 TO 7, RUN THIS PROMPT MULTIPLE TIMES, THEN COMBINE THE OUTPUTS IN A FINAL PASS.] The research question I was trying to answer is: [ONE SENTENCE] Produce a synthesis in this exact format: 1. TOP 5 THEMES: ranked by frequency. For each theme, state the theme, the number of interviewees who raised it, and one representative quote or paraphrase. 2. CONTRADICTIONS: places where interviewees disagreed with each other, or where their stated desires contradict their reported behaviour. Name the contradiction, the segments on each side, and the most likely reason. 3. SURPRISING FINDINGS (3 to 5): things that were not in the research question but emerged anyway. 4. WHAT WAS NOT SAID: topics you would have expected interviewees to mention that were absent. Speculate on why. 5. SEGMENT DIFFERENCES: where themes broke down differently by customer segment, role, geography, or maturity. 6. RESEARCH GAPS: the three questions this set of interviews could not answer, that the next round should. Output format: numbered sections 1 to 6 in the order above. Constraints: - Do not cherry-pick to support a pre-existing hypothesis. If the data contradicts the research question, say so. - Do not invent quotes. If a quote is paraphrased, label it "(paraphrase)". - Do not collapse contradictions into a single theme. Preserve disagreement. - If fewer than 5 themes are warranted by the data, return fewer. - Use plain English. Do not translate customer language into marketing jargon.
Usage Notes
Use on Day 2. Most LLMs degrade past around 30,000 tokens of pasted notes, so batch above ten interviews. The "contradictions" and "what was not said" sections are usually where the strategic insight lives.
Worked example shape
Ten interviews with mid-market CFOs surface themes around forecast uncertainty, regulatory burden, and CFO-CEO trust. Contradictions appear between manufacturing and services CFOs on cash-flow visibility tools. Surprising finding: half raised AI literacy unprompted.
The Option Generator
Purpose: Produces a list of 8 to 12 strategic options ranging from incremental to ambitious. Forces breadth before narrowing.
Copy and Paste
The strategic question I am trying to answer is: [ONE SENTENCE: THE QUESTION, NOT THE ANSWER] Context the model needs: [PASTE: BUSINESS MODEL, CURRENT STAGE, RESOURCE CONSTRAINTS, KEY CUSTOMER SEGMENTS, AND ANY INSIGHTS ALREADY GENERATED FROM market-size / competitive-map / customer-synthesis] Primary constraint to respect: [INSERT: CAPITAL, TIME, TALENT, REGULATORY, OR OTHER] Produce a list of strategic options in this exact format: 1. INCREMENTAL OPTIONS (3 to 4): adjustments to the current strategy. Lower risk, lower upside. 2. ADJACENT OPTIONS (3 to 4): meaningful pivots that build on existing capabilities. Medium risk, medium upside. 3. AMBITIOUS OPTIONS (2 to 3): substantial reorientations of the business. High risk, high upside. 4. ANTI-OPTIONS (1 to 2): strategies that look attractive but are actively wrong for this business. Name them so they are eliminated rather than re-litigated later. For each option, include: - A one-line description. - The primary bet being made (what must be true for it to work). - The primary risk. - A rough resource implication: Low, Medium, or High. Output format: tiered sections 1 to 4 with the four-line structure per option. Constraints: - Do not list only incremental options. If the ambitious tier feels uncomfortable, that is the signal it is the right tier. - Do not list more than 12 options total. Beyond 12 the exercise becomes brainstorming, not strategy. - Do not include options that are obviously identical to current strategy. The current strategy is the baseline, not an option. - Each option must be distinct. If two options share the same primary bet, merge them. - Respect the primary constraint stated above. An option that requires capital we do not have is not an option.
Usage Notes
Use on Day 3. The discomfort with the ambitious tier is the diagnostic. If the ambitious options feel safe, push the model harder. Anti-options prevent the team from re-debating obviously-wrong directions throughout the sprint.
Worked example shape
For a 20-person AI consultancy under a "no incremental headcount" constraint: incremental options include pricing changes and service-line tweaks, adjacent options include productising a service, ambitious options include shutting consulting to build a product company.
The Strategy Choice
Purpose: Forces an explicit, defensible strategic choice in a form that can be communicated to the board and the team without loss of fidelity.
Copy and Paste
We have generated the following strategic options: [PASTE OPTIONS LIST FROM PRIOR PROMPT] The criteria we are deciding on are (ranked): [PASTE CRITERIA FROM DECISION FRAME] Produce a strategic choice document in this exact format: 1. THE CHOICE (one sentence): "We will [DO X] and not [DO Y], because [PRIMARY REASON]." 2. WHAT WE ARE DOING (3 to 5 bullet points): the specific actions, investments, and bets we are making. 3. WHAT WE ARE NOT DOING (3 to 5 bullet points): the alternatives we are explicitly declining and the reasons for declining each. 4. WHY THIS BEATS THE ALTERNATIVES (paragraph form): a defensible argument referencing the ranked criteria. 5. SUCCESS MEASURES (3 to 5 bullet points): the specific outcomes that will tell us in 12 months whether the choice was correct. 6. STOP CONDITIONS (3 bullet points): the specific signals that would tell us to reverse or pivot. Output format: numbered sections 1 to 6 in the order above. Constraints: - Do not hedge. Strategic choices that hedge are not strategic choices. - Do not list more than 5 items in any category. Strategic clarity requires compression. - If the options provided do not support a confident choice, say so. Do not manufacture confidence.
Usage Notes
Use on Day 4, after option generation and before narrative crafting. The "we will not" list is what makes the strategy real. Most strategy outputs list what the team will do and skip what they will not. If the model returns a vague "we will do everything" output, the input options are not yet differentiated enough; return to prompt 4.
Worked example shape
The output reads in the shape of "We will focus on UK enterprise mid-market and not pursue SMB or US, because mid-market matches our deal-size and sales-motion economics", followed by 4 doing-bullets, 4 not-doing bullets, paragraph rationale, 4 success measures, 3 stop conditions.
The Narrative Crafter
Purpose: Translates the chosen strategy into a one-page narrative readable by the board, the team, and the customer.
Copy and Paste
The strategic choice we have made is:
[PASTE OUTPUT OF strategy-choice PROMPT]
The audiences for this narrative are:
[LIST: BOARD / EXECUTIVE TEAM / WHOLE COMPANY / CUSTOMERS / INVESTORS / PUBLIC]
The existing narrative the audiences are operating under (if any):
[PASTE EXISTING POSITIONING OR WRITE "NONE"]
Produce a one-page strategy narrative in this exact format:
1. THE OPENING (2 to 3 sentences): the world has changed in a specific way, and that change creates a specific opportunity or threat.
2. THE PROBLEM WE ARE SOLVING (one paragraph): the customer pain or market gap, named concretely.
3. WHAT WE ARE DOING ABOUT IT (one paragraph): the strategic choice in plain English, not in consulting jargon.
4. WHY US, WHY NOW (one paragraph): the specific reasons this business is positioned to win, and why the timing is right.
5. WHAT SUCCESS LOOKS LIKE IN 12 MONTHS (3 to 4 bullet points): the concrete outcomes a reader can check us against.
6. WHAT WE ARE NOT DOING (2 to 3 sentences): explicit rejection of the most attractive alternatives, to pre-empt board or team second-guessing.
Output format: numbered sections 1 to 6 in the order above.
Constraints:
- Do not use consulting jargon. If a sentence could appear in any company's strategy doc, rewrite it.
- Do not exceed 500 words total.
- Do not soften the "what we are not doing" section. Specificity is the point.
- Match the voice of the named audience. For the whole company, use first-person plural. For customers, address them directly.
- Do not include corporate filler ("synergies", "ecosystem", "best-in-class", "world-class").
Usage Notes
Use on Day 4. The output should be readable aloud by a non-strategist in two minutes. If you cannot read it aloud without stumbling, rewrite.
Worked example shape
A 400-word board narrative opens with a specific market change ("AI procurement has shifted from experiment to budget line in the last 12 months"), states the strategic focus on enterprise mid-market, names three 12-month outcomes, and closes with an explicit "not pursuing SMB" line.
The Decision Log
Purpose: Records the strategic decision, the rationale, the alternatives considered, the reversibility, and the success measures.
Copy and Paste
The strategic decision being logged is: [PASTE OUTPUT OF strategy-choice PROMPT, OR DESCRIBE THE DECISION IN ONE PARAGRAPH] Date of decision: [DATE] Decision-makers: [NAMES AND ROLES] Decision-making forum: [BOARD / EXECUTIVE TEAM / FOUNDER ALONE / OTHER] Produce a decision log entry in this exact format: 1. DECISION (one sentence). 2. CONTEXT (one paragraph): the situation that made this decision necessary now. 3. ALTERNATIVES CONSIDERED (3 to 5): each named with one line on why it was rejected. 4. RATIONALE (paragraph): the primary reasoning, referenced to evidence where available. 5. REVERSIBILITY: classify as Type 1 (one-way door, hard to reverse) or Type 2 (two-way door, low cost to reverse). State the reasoning. 6. SUCCESS MEASURES (3 to 5 bullet points): the metrics that will tell us in 6 and 12 months whether the decision was right. 7. REVIEW DATE: the date this decision should be revisited regardless of metrics. 8. DISSENT RECORDED: any decision-maker who disagreed, with their reasoning preserved. Output format: numbered sections 1 to 8 in the order above. Constraints: - Do not rewrite history. The log should reflect the reasoning at the time of decision, even if subsequent evidence shifts. - Do not omit dissent. If a decision-maker disagreed, record their position with the same care as the consensus position. - Do not classify every decision as Type 1. Most strategic decisions are Type 2; over-classifying as Type 1 produces decision paralysis. - If no alternatives were considered, say so. That is itself a finding.
Usage Notes
Use on Day 5. The Type 1 / Type 2 reversibility framing (one-way door / two-way door, popularised by Jeff Bezos) is the single most useful field. The dissent line is non-optional; quiet dissent corrodes commitment.
Worked example shape
A log for a decision to focus on UK enterprise mid-market classifies the decision as Type 2 (reversible within 6 months at modest cost), records that one co-founder preferred the adjacent option, and sets a review date 12 months out tied to a specific milestone.
The 90-Day Execution Plan
Purpose: Translates the chosen strategy into a 90-day plan with named owners, weekly milestones, and explicit stop conditions.
Copy and Paste
The strategic choice we are executing is: [PASTE OUTPUT OF strategy-choice PROMPT] The team available to execute is: [PASTE TEAM ROSTER WITH ROLES, AVAILABILITY, AND ANY KEY CONSTRAINTS] Budget envelope for the 90 days: [STATE BUDGET, OR WRITE "TIME-ONLY, NO INCREMENTAL BUDGET"] Produce a 90-day execution plan in this exact format: 1. THE 90-DAY OBJECTIVE (one sentence): the one outcome that, if achieved, proves the strategy is working. 2. WORKSTREAMS (3 to 5): for each, name the workstream, the named owner from the roster, the 90-day deliverable, and the budget allocation. 3. WEEKLY MILESTONES (13 weeks): for each week, one or two specific milestones. Each milestone must be observable, meaning a thing that exists at end of week. Not a status. 4. DEPENDENCIES: cross-workstream dependencies with the week they become critical. 5. STOP CONDITIONS (3 to 5): the specific trigger metric, threshold, and action. Format each as "If [METRIC] [CONDITION] [VALUE] by [WEEK], then [ACTION]." 6. WHAT IS NOT IN THIS PLAN (3 to 5 bullet points): the activities the team will explicitly not do during these 90 days, to protect focus. 7. REVIEW CADENCE: the weekly, fortnightly, or monthly forum that will review progress, and who attends. Output format: numbered sections 1 to 7 in the order above. Constraints: - Do not assign workstreams to "the team" or "marketing". Every workstream has a named owner from the roster I provided. - Do not list a milestone that cannot be observed at end of week. "Make progress on hiring" is not a milestone. "First-round interviews completed with 4 candidates" is a milestone. - Do not omit stop conditions. A plan with no stop conditions is a plan that cannot be falsified. - If the team available is insufficient for the workstreams identified, say so and recommend a smaller set.
Usage Notes
Use on Day 5 as the closing artefact of the sprint. The "what is not in this plan" list serves the same function as the "we will not" list in the strategy choice; it protects focus.
Worked example shape
For a UK probate-tech startup post-strategy, the 90-day objective might be "10 paying enterprise accounts in mid-market law firms", workstreams cover sales, product, and partnerships, weekly milestones name specific deliverables (a contract signed, a feature shipped, a partnership announced), and stop conditions include lines like "If conversion rate below 8% by week 8, pause outbound and re-segment."
How to run the pack
These eight prompts are working tools, not inspiration. They are designed to force constraints, highlight tradeoffs, and prevent the most common failure modes of strategic planning. For best results, run them sequentially over a five-day period with your core team. Run them in one focused day with two or three people if the question is urgent.
One pack rule: if any prompt returns a hedged or vague output, return to the previous prompt and tighten the inputs. The discipline is in the inputs, not the prompts.
No affiliate links included in this issue.
|
Zymbos Intelligence
zymbos.ai
You are receiving this because you subscribed at zymbos.ai
© 2026 Zymbos Intelligence · John McGann · London, UK Zymbos Ltd · Company No. 16198848 · Teddington, England |