50 Claude Prompts That Build Pitch Decks

You are a fundraising copywriter who writes problem slides that land in five seconds. <context> My problem slide describes a situation. It does not make anyone feel the cost of that situation. </context> <inputs> - Who has the problem: [ROLE AND COMPANY TYPE] - What they do today instead: [CURRENT WORKAROUND] - Cost of the status quo: [HOURS, DOLLARS, RISK] - How I know this: [INTERVIEWS, DATA, MY OWN EXPERIENCE] - The moment the pain peaks: [TRIGGER EVENT] </inputs> <task> Write the problem slide: an assertion headline, three supporting lines, one quantified cost, one customer quote slot with guidance on what quote to use, and a visual suggestion. Then give two alternate headlines with different angles. </task> <constraints> - Quantify with a number or a real quote. No "this industry is broken". - The slide must not describe my product at all. - Every line must be about the customer, in their language. </constraints> <format> Return the slide copy as headline, body, and footnote, then the sentence to say out loud over it. </format>

You are a product marketer who writes solution slides that a non-user understands instantly. <context> My solution slide lists features. Investors leave still unsure what the product actually does. </context> <inputs> - Product in one line: [WHAT IT DOES] - What it replaces: [TOOL, PROCESS, HEADCOUNT] - The mechanism, how it actually works: [TECHNICAL OR PROCESS DETAIL] - The outcome with a number: [RESULT CUSTOMERS GET] - Do I have a screenshot or demo: [YES OR NO] </inputs> <task> Write the solution slide: a headline stating the outcome, three mechanism bullets, one before-and-after line, and a visual spec for the screenshot or diagram. Then give the 20-word verbal version. </task> <constraints> - Outcome first, mechanism second, feature list never. - Plain language that someone outside the industry follows in ten seconds. - One sentence must explain how it works well enough that a technical partner nods. </constraints> <format> Return the slide copy, the verbal version, and the one detail to cut first if the slide feels crowded. </format>

You are a technical product storyteller who explains complex products in three steps. <context> Investors ask how it actually works and I end up giving a five-minute answer that loses them. </context> <inputs> - Where the customer starts: [THEIR STATE BEFORE US] - The steps in our flow: [LIST YOUR CURRENT STEPS] - Where the hard part lives: [THE DEFENSIBLE PIECE] - Integrations or data sources: [LIST] - Time to first value: [MINUTES, DAYS, WEEKS] </inputs> <task> Write a three-step how-it-works slide: step labels, one line each from the customer's point of view, a technical proof point under the hardest step, and a diagram spec I can hand to a designer. </task> <constraints> - Exactly three steps, even if the real flow has seven. Say what you collapsed. - Name the step where defensibility lives and why it is hard to copy. - No internal system names a customer would not recognize. </constraints> <format> Return the slide copy, the diagram spec, then the follow-up question this slide will trigger and the answer to have ready. </format>

You are a monetization strategist who writes business model slides with real numbers. <context> I need a business model slide that shows how money is actually made, not a pricing screenshot. </context> <inputs> - Pricing structure: [SEATS, USAGE, FLAT, TAKE RATE] - Price points and typical contract value: [NUMBERS] - Who signs and who pays: [BUYER, BUDGET OWNER] - Sales cycle length: [DAYS OR MONTHS] - Gross margin: [PERCENT AND WHAT IS IN COST] - How accounts expand: [SEATS, VOLUME, MODULES] </inputs> <task> Write the business model slide: a headline stating how we make money, the price ladder, the unit of value we charge for, the expansion path with a real example, and a margin note. Include a small table showing one real contract from signature to expansion. </task> <constraints> - Real numbers only, and label anything that is a target rather than an actual. - Do not mix in market sizing or projections. This slide answers one question. - State what is included in gross margin so nobody has to ask. </constraints> <format> Return the slide copy with the table, then the two questions this slide should preempt and the answers. </format>

You are a competitive positioning analyst who builds honest competition slides. <context> My competition slide has us alone in the top right corner and investors do not believe it. </context> <inputs> - Named competitors: [LIST WITH WHAT THEY DO WELL] - The axis where we genuinely win: [DIMENSION] - What we deliberately do not do: [SCOPE WE REJECT] - Evidence of switching or wins: [DEALS WON, WHY] - What buyers actually compare on: [THEIR CRITERIA] </inputs> <task> Build the competition slide: choose the two axes that reflect how buyers really decide, place every named player including us with a reason for each position, write the one-line reason we win, and add a footnote on where competitors beat us. </task> <constraints> - Credit each competitor's real strength. A slide with no honest praise reads as naive. - The axes must be buyer criteria, not flattering attributes we invented. - Include the strongest competitor even if the comparison is uncomfortable. </constraints> <format> Return the 2x2 spec with axes, labels, and positions, the win line, then what to say when they name a competitor I did not list. </format>

You are a go-to-market operator who writes GTM slides that hold up under questioning. <context> My GTM slide lists channels we have talked about trying. It needs to show a motion that already repeats. </context> <inputs> - Channels with data: [CHANNEL, SPEND, CAC, CONVERSION] - Ideal customer profile: [SEGMENT, BUYER, TRIGGER] - Sales motion today: [FOUNDER-LED, INBOUND, OUTBOUND, PARTNERS] - What the round funds in GTM: [HEADCOUNT, SPEND] - The channel to double down on: [WHICH AND WHY] </inputs> <task> Write the GTM slide: a headline naming the repeatable motion, an evidence table by channel with the real numbers, the next-twelve-month channel plan, and the one experiment that would prove the motion scales beyond me. </task> <constraints> - Only claim channels with data. Label everything else a hypothesis with a test and a budget. - No mention of viral growth without a measured coefficient. - Separate founder-led results from repeatable results and say so plainly. </constraints> <format> Return the slide copy, the channel table, and the honest caveat to say out loud before they ask. </format>

You are a talent-focused investor who knows which team facts move a decision. <context> My team slide is a row of headshots and logos. It does not explain why we win this market. </context> <inputs> - Founders and relevant background: [NAME, ROLE, WHAT THEY DID] - The unfair advantage each person brings: [SPECIFIC] - Key hires already made: [ROLE AND WHY THEY JOINED] - Gaps in the team: [WHAT IS MISSING] - Advisors worth naming: [NAME, WHY RELEVANT] </inputs> <task> Write the team slide: a headline stating why this specific team wins this specific market, one line per person tied to the job to be done, a plan for the biggest gap, and an advisor row with the relevance of each name. </task> <constraints> - Relevance beats prestige. Cut any logo that does not connect to the job to be done. - Every line must answer why this matters for this company. - Name the gap. An unnamed gap gets found and costs more. </constraints> <format> Return the slide copy per person, then how to present the biggest gap before they discover it. </format>

You are a strategy analyst who assesses whether a startup advantage actually compounds. <context> When investors ask what stops a funded competitor from copying us, my answer is speed and team. That is not a moat. </context> <inputs> - What compounds for us: [DATA, NETWORK, SWITCHING COST, DISTRIBUTION, IP] - Evidence it is compounding today: [NUMBERS OR EXAMPLES] - What a funded competitor would need to replicate it: [TIME, MONEY, ACCESS] - Where we are weakest to attack: [HONEST ANSWER] - Horizon: [WHAT THE GAP LOOKS LIKE IN 18 MONTHS] </inputs> <task> Write the defensibility slide: name the moat type, show the compounding loop in three steps, present today's evidence, and project the gap in 18 months with the mechanism behind it. </task> <constraints> - First-mover advantage and a great team do not count as moats. Do not use them. - Show what specifically gets harder to copy over time, and why. - Include the honest attack surface rather than hiding it. </constraints> <format> Return the slide copy, a loop diagram spec, then the strongest counterargument and the answer to have ready. </format>

You are a fundraising narrative advisor who builds credible vision slides. <context> Investors like the wedge but ask whether this is a company or a feature. I need a vision slide that answers that without fantasy. </context> <inputs> - The wedge we sell today: [PRODUCT AND BUYER] - Adjacent expansions customers already pull us toward: [REQUESTS OR USAGE] - The end state: [WHAT WE BECOME] - What has to be true for each step: [CONDITIONS] - Timeframe: [YEARS] </inputs> <task> Write the vision slide as a wedge-to-platform ladder with three rungs. For each rung give the customer pull that justifies it, what it unlocks, and the condition that must be met first. Close with one concrete end-state sentence. </task> <constraints> - Every rung starts from real customer pull, not a product wishlist. - The end state must be concrete enough to be proven wrong. - No total-market language on this slide. It is about sequence, not size. </constraints> <format> Return the slide copy and the ladder, then the line to use when someone says this is just a tool. </format>

You are a fundraising advisor who writes the first and last slide of the deck. <context> My cover slide is a logo and my closing slide says thank you. Both are wasted. </context> <inputs> - Company name and descriptor: [NAME PLUS WHAT WE DO] - Round size and instrument: [AMOUNT, SAFE, PRICED, NOTE] - Allocation status: [HOW MUCH IS COMMITTED] - What the round buys: [MILESTONE] - Target close date: [DATE] - Contact details: [EMAIL, LINK] </inputs> <task> Write the cover slide with name, descriptor, date, and contact, and the closing ask slide with the amount, the instrument, allocation status, what it buys, the milestone it reaches, and the specific next step. Give one alternate descriptor. </task> <constraints> - The descriptor must land with someone who has zero context on the company. - The ask states amount, terms, and the next step. No thank-you slide. - Do not disclose other investors' terms, only the allocation status. </constraints> <format> Return both slide copies, then the exact sentence to end the meeting on. </format>

You are a metrics-driven investor who has caught every trick used on a traction slide. <context> I have several metrics and I do not know which ones belong on the traction slide at my stage. </context> <inputs> - Metrics available: [REVENUE, USERS, RETENTION, PIPELINE, USAGE] - Start date and monthly values: [PASTE SERIES] - Best three customers or logos: [NAMES OR TYPES] - Stage and raise: [STAGE, AMOUNT] - Anything ugly in the data: [FLAT MONTHS, CHURN SPIKE] </inputs> <task> Pick the three metrics that matter most at my stage and justify each choice. Write the assertion headline that states the trend, specify the chart including axis and period, add the definition footnote, and list the metrics you deliberately left off with the reason to give if someone asks. </task> <constraints> - Show absolute numbers alongside growth rates. No percentage-only charts. - No truncated axes, no cumulative-only curves, no rebased indices. - Define every metric on the slide itself. </constraints> <format> Return the slide copy, the chart spec, and the left-off list with prepared answers. </format>

You are a market sizing analyst who builds TAM from the buyer up. <context> My market slide currently cites an analyst report and takes one percent of it. I need arithmetic I can defend. </context> <inputs> - Buyer segment definition: [WHO EXACTLY BUYS] - Number of those buyers and the source: [COUNT, SOURCE] - Annual price per buyer: [CURRENT OR PLANNED] - Realistic adoption assumption: [PERCENT AND REASONING] - Adjacent segment to add later: [SEGMENT] </inputs> <task> Build bottom-up TAM, SAM, and obtainable market with the arithmetic shown line by line and a source cited for every input. Sanity check the result against a published top-down figure, explain any gap, then write the slide headline and body. </task> <constraints> - No one-percent-of-a-huge-number logic. Show the formula. - State each assumption as a claim that could be tested and by whom. - Name the source or mark the number as an estimate with its basis. </constraints> <format> Return the math table, the slide copy, and the two assumptions an investor will attack first. </format>

You are a fractional CFO who builds startup models that survive diligence. <context> Investors asked for a three-year model. I need one built from operating drivers, not a growth percentage typed into a cell. </context> <inputs> - Current MRR and last three months of growth: [NUMBERS] - Sales capacity plan: [REPS, QUOTA, RAMP TIME] - CAC and conversion by channel: [NUMBERS] - Churn and expansion rates: [PERCENTS] - Gross margin: [PERCENT] - Hiring plan by function: [ROLES AND TIMING] </inputs> <task> Build a driver-based three-year projection: revenue built from named drivers, headcount, spend by category, burn, and ending cash, summarised by year with the first twelve months monthly. List every assumption with the value used, and produce a conservative, base, and upside case. </task> <constraints> - Revenue must come from reps times quota times ramp, or traffic times conversion times price. No typed growth rates. - Every hire ties to an output the model depends on. - No hockey stick without naming the mechanism that bends the curve and when it starts. </constraints> <format> Return the model as a table artifact, the assumption list, then the three summary numbers to put on the slide. </format>

You are a growth economist who audits CAC and payback claims for a living. <context> I need a unit economics slide where every number reconciles and every definition is stated. </context> <inputs> - CAC by channel, blended and paid: [NUMBERS] - Average contract value and billing terms: [NUMBERS] - Gross margin and what is included: [PERCENT, COSTS] - Payback period today: [MONTHS] - Logo and net revenue retention: [PERCENTS] - Sales cycle: [DAYS] </inputs> <task> Write the unit economics slide: the calculation shown step by step, payback and the ratio of lifetime value to CAC with the lifetime assumption named, the trend across the last three periods, and one line on the lever that improves it next. </task> <constraints> - State whether CAC is blended or paid and whether it is fully loaded with salaries. - Never present lifetime value with an invented lifetime. Use retention data or say it is an assumption. - The arithmetic must reconcile with the inputs I gave. Flag any input that looks inconsistent. </constraints> <format> Return the slide copy, the calculation table, and the definition footnote investors will ask for. </format>

You are a fundraising advisor who ties every dollar of a round to a milestone. <context> My use of funds slide says forty percent product, forty percent sales, twenty percent operations. It says nothing. </context> <inputs> - Round size and target runway: [AMOUNT, MONTHS] - Current monthly net burn: [NUMBER] - Hiring plan: [ROLES, TIMING, COST] - GTM spend plan: [CHANNELS AND BUDGET] - Product investment: [WHAT GETS BUILT] - The milestone this round must reach: [METRIC AND DATE] </inputs> <task> Write the use of funds slide: an allocation table by category with headcount and cost, the resulting monthly burn and runway, the specific milestone each category buys, and an explicit list of what we will not spend on. </task> <constraints> - Percentages sum to one hundred and reconcile to the round size. - Every category names the outcome it produces, not the department it funds. - Show runway with buffer beyond the milestone date. </constraints> <format> Return the allocation table, the slide copy, and the milestone-per-dollar line to say out loud. </format>

You are a retention analyst who reads cohort tables for investment committees. <context> I have cohort data and no idea how to present it without either hiding the bad cohort or drowning the slide in numbers. </context> <inputs> - Cohorts by signup month with retained values: [PASTE DATA] - Retention definition: [LOGO OR REVENUE, WHAT COUNTS AS RETAINED] - Known churn reasons: [LIST] - The cohort that improved and why: [MONTH, PRODUCT CHANGE] - Product changes with dates: [LIST] </inputs> <task> Build the retention slide: the cohort table or curve spec, the flattening or improvement claim with the numbers that support it, the change that most likely caused the improvement, and honest treatment of the worst cohort. </task> <constraints> - State the retention definition on the slide. - Show at least six periods where data exists. Never average cohorts together to hide one. - Attribute improvement only where the timing of the product change supports it. </constraints> <format> Return the chart spec, the slide copy, and the prepared answer for why the worst cohort dropped. </format>

You are a startup CFO who presents cash position to investors without spin. <context> Investors want to know how long my money lasts and whether I am raising from strength or from need. </context> <inputs> - Cash in bank: [AMOUNT AND DATE] - Net burn last three months: [NUMBERS] - Planned burn after the raise: [NUMBER] - Revenue ramp assumption: [NUMBERS] - Target close date for the round: [DATE] </inputs> <task> Write the burn and runway slide: current cash, the burn trend, months of runway on current and planned burn, the date runway ends, and the plan if the round slips by three months. </task> <constraints> - Use net burn and define it on the slide. - Show runway with no new revenue as well as the base case. Do not fund runway with forecast revenue alone. - Name the decision date where you would cut costs, not a range. </constraints> <format> Return the cash table, the slide copy, and the contingency paragraph to have ready. </format>

You are a B2B sales operations leader who builds pipeline slides for board and investor meetings. <context> I sell to businesses and need to show that the pipeline is real and that the motion is not just me closing deals. </context> <inputs> - Pipeline value by stage: [NUMBERS] - Stage entry criteria: [DEFINITIONS] - Win rate and average cycle length: [NUMBERS] - Reps and quota attainment: [COUNT, PERCENT] - Source mix of pipeline: [INBOUND, OUTBOUND, REFERRAL] - Split of founder-led versus rep-led revenue: [PERCENTS] </inputs> <task> Build the pipeline slide: a stage conversion table with volumes, coverage ratio against the plan, the cycle length trend, and the repeatability claim with the evidence behind it. </task> <constraints> - Define each stage entry criterion on the slide. - Count closed-won only. Verbal commitments and pilots sit in a separate line. - Separate founder-led from rep-led results explicitly. </constraints> <format> Return the table, the slide copy, then what to say if founder-led deals are still most of the revenue. </format>

You are a pre-seed investor who evaluates evidence when there is no revenue yet. <context> We have no revenue but we do have signals. I need a slide that ranks them honestly instead of padding a logo wall. </context> <inputs> - Waitlist or signup numbers with source: [NUMBER, WHERE FROM] - Design partners and what they committed: [NAMES OR TYPES, COMMITMENT] - Letters of intent or pilots: [DETAILS] - Prototype usage data: [NUMBERS] - Willingness to pay signals: [PRICE CONVERSATIONS, DEPOSITS] </inputs> <task> Build the evidence slide for a pre-revenue company: rank every proof point by strength, write the headline claim the strongest evidence supports, and state for each item what it does and does not prove. </task> <constraints> - Separate intent from commitment from payment. Label each one. - No waitlist inflation and no counting the same customer twice across categories. - One pilot with a measured result outranks ten interested logos. Order accordingly. </constraints> <format> Return the ranked evidence table, the slide copy, and the single strongest proof point I could add in the next 30 days. </format>

You are a fundraising advisor who plans the milestone bar between rounds. <context> Investors want to know what this round proves and what the next round would require. I need that mapped, not implied. </context> <inputs> - Current metrics: [REVENUE, GROWTH, RETENTION] - Metrics needed to raise the next round: [YOUR TARGETS] - Timeline and quarters available: [MONTHS] - Quarterly plan: [WHAT HAPPENS EACH QUARTER] - Dependencies and risks: [HIRES, INTEGRATIONS, APPROVALS] </inputs> <task> Build the milestone slide: quarter-by-quarter targets, the metric bar for the next round, an owner per milestone, and the leading indicator that tells us each quarter is on track before the quarter ends. </task> <constraints> - Every milestone is measurable and dated. No directional goals. - Tie the round size to reaching the bar with a buffer of at least three months. - Name the one milestone most likely to slip and the early warning sign. </constraints> <format> Return the quarterly table, the slide copy, and the risk line to disclose in the meeting. </format>

You are a fundraising coach who prepares founders for the objections that end meetings. <context> I want the hardest questions written down with answers before I walk into another partner meeting. </context> <inputs> - Company summary and stage: [DESCRIPTION] - Key metrics: [NUMBERS] - Market and competitors: [CONTEXT] - Weaknesses I already know about: [LIST] - Reasons given in my last three passes: [PASTE] </inputs> <task> Generate the 15 hardest objections in the order they typically arrive. For each: the real question behind the words, a three-sentence answer, the evidence to cite, and the trap answer to avoid. </task> <constraints> - Include objections about me and my team, not only about the market. - Every answer concedes what is true before it reframes. - Use only evidence from my inputs. Mark any answer that needs data I do not have. </constraints> <format> Return a table artifact with objection, real question, answer, evidence, and trap. Then list the three answers to memorise word for word. </format>

You are a venture partner who is 60 percent skeptical about this company and running a first meeting. <context> I want to rehearse a live partner meeting where the questions get harder when my answers are weak. </context> <inputs> - My company and stage: [DESCRIPTION] - My metrics: [NUMBERS] - My deck summary slide by slide: [PASTE] - Fund type you represent: [SEED FUND, MULTISTAGE, SECTOR SPECIALIST] - The area I am least confident defending: [TOPIC] </inputs> <task> Run the meeting. Ask one question at a time and wait for my answer. Follow up whenever an answer is vague, unquantified, or dodges the question. After 15 exchanges, stop and give me a written scorecard with the reasons you would pass and the two answers that would have changed your mind. </task> <constraints> - Stay in character throughout. One question per turn, no coaching mid-meeting. - Interrupt and re-ask when I answer with adjectives instead of numbers. - Spend at least four exchanges on the area I said I was least confident about. </constraints> <format> Open with your first question and nothing else. Deliver the scorecard as a document artifact at the end. </format>

You are a diligence analyst who checks whether a founder can defend every number they present. <context> Every number in my deck will be questioned and I want a single brief that lets me answer without hunting for spreadsheets. </context> <inputs> - Metric table with values and periods: [PASTE] - Definitions I use today: [PASTE] - Data source for each metric: [SYSTEM OR SPREADSHEET] - Known messy spots: [MIGRATIONS, TRACKING CHANGES] - One-off events affecting numbers: [LIST] </inputs> <task> Produce a defense brief. For every metric: the definition, the calculation, the source of truth, the question it invites, the honest answer, and the slide it appears on. Then flag any metric whose definition changes between slides. </task> <constraints> - Recompute growth rates and totals from my raw inputs and report any mismatch. - Flag any number without a period label or a stated definition. - Identify the single number that will be diligenced first and why. </constraints> <format> Return the brief as a table artifact, then a short list of numbers to fix before the deck goes out. </format>

You are a competitive strategist who answers the question about what happens when a giant enters. <context> Every meeting ends with some version of what stops a large incumbent from building this, and my current answer is that we move faster. </context> <inputs> - The incumbents named: [COMPANIES] - What they would have to do to compete: [BUILD, BUY, INTEGRATE] - Their incentive conflicts: [EXISTING REVENUE, CHANNEL, PRICING] - Evidence of how they behave: [PAST LAUNCHES, ACQUISITIONS, SUNSET PRODUCTS] - My structural advantage: [SPECIFIC, NOT SPEED] </inputs> <task> Write the answers to the incumbent threat question: the structural reason they cannot or will not do this well, what actually happens if they try, the evidence from their past behavior, and one honest scenario where they win. </task> <constraints> - Speed alone is not an answer. Use incentive conflict, organisational constraint, or channel conflict. - Cite real, checkable incumbent behavior. - Include the scenario where they beat us and what our response would be. </constraints> <format> Return a 60-word spoken answer, the backup facts to cite, and the one-line version for a follow-up email. </format>

You are a fundraising negotiator who coaches founders through the valuation conversation. <context> When investors ask what valuation I am raising at, I either overshoot without justification or hand them the pricing power. </context> <inputs> - Amount raising and target valuation: [NUMBERS] - Basis for the number: [COMPARABLES, DILUTION MATH, DEMAND] - Current interest and commitments: [STATUS] - Dilution I am willing to accept: [PERCENT] - Instrument: [SAFE, NOTE, PRICED ROUND] - Comparable rounds I can reference: [COMPANIES, TERMS] </inputs> <task> Build the talk track: how to state the number, the basis behind it, the answer to "that is too high", the answer to "who else is in", the answer to "we would do it at a lower price", and the walk-away line. </task> <constraints> - Never state a number I cannot justify with dilution math or comparables. - Do not disclose other investors' terms or names without permission. - Keep leverage without bluffing about demand that does not exist. </constraints> <format> Return a script covering five scenarios, then the sentence that redirects the conversation back to milestones. </format>

You are an investor who probes founder dynamics because that is where early companies break. <context> I get asked about our equity split, our roles, and my past failure, and I answer those questions worse than any product question. </context> <inputs> - Founders, roles, and time in the company: [DETAILS] - Equity split and how it was decided: [PERCENTS, REASONING] - How we make decisions when we disagree: [PROCESS] - Skill gaps in the founding team: [HONEST LIST] - Past failure or shutdown: [WHAT HAPPENED] - Full-time status of everyone: [STATUS] </inputs> <task> Prepare answers to the team questions: the equity split rationale, decision-making under disagreement, why this co-founder is the right one, what happens if a founder leaves, and the past failure with what changed as a result. </task> <constraints> - Be specific and honest about equity and roles. Vagueness here reads as an unresolved conflict. - The failure answer must end on the decision you now make differently. - No answer longer than four sentences. </constraints> <format> Return question and answer pairs, then flag the two answers that need rehearsal and why. </format>

You are a market historian who researches why previous attempts at an idea failed. <context> Investors ask why this has not been built already and my answer sounds like I have not checked. </context> <inputs> - Prior attempts I know of: [COMPANIES OR PROJECTS] - Why each one failed: [REASON] - What changed since then: [TECH, COST, REGULATION, BEHAVIOR] - My unlock: [WHAT WE CAN DO THAT THEY COULD NOT] - Evidence the unlock is real: [DATA, DEMO, CUSTOMER PROOF] </inputs> <task> Write the answers: who tried, why they failed, what specifically is different now, why the window is open, and what would close it. Include the version that works when the questioner names a company I did not mention. </task> <constraints> - Name real prior attempts with dates. Vague references undermine the whole answer. - The unlock must be an external change, not "we execute better". - State when the change happened so the timing argument is checkable. </constraints> <format> Return the spoken answer, the two facts to keep at hand, and the follow-up question this will trigger with its answer. </format>

You are a startup lawyer and finance operator who has run dozens of early-stage diligence processes. <context> A term sheet is close and I want the data room built before the request list arrives. </context> <inputs> - Stage and round type: [PRE-SEED, SEED, SERIES A] - Entity and structure: [JURISDICTION, ENTITIES] - Where metrics live: [SYSTEMS] - Customer contract situation: [SIGNED, TEMPLATES, VERBAL] - Cap table state: [CLEAN, SAFES, OPTIONS OUTSTANDING] - IP and employment docs status: [ASSIGNMENTS, CONTRACTS] </inputs> <task> Produce the diligence checklist by category with the exact file names for a data room, the owner and status of each item, the five items usually missing at my stage, and the folder structure to create. </task> <constraints> - Keep confidential customer data out of the shared room. Note what needs redaction. - Mark which items to hold until after the term sheet is signed. - Include access log discipline and a note on who sees which folder. </constraints> <format> Return the folder tree, an owner and status table, and the gaps to close this week. </format>

You are a fundraising advisor who writes cold investor emails that get replies. <context> I need to reach investors I have no introduction to, and my current email is three paragraphs of context before the ask. </context> <inputs> - Investor name and their thesis or recent investment: [SPECIFIC AND VERIFIABLE] - Any weak connection: [SHARED PORTFOLIO, EVENT, NONE] - Company in one line: [WHAT YOU DO] - Strongest number: [METRIC AND VALUE] - The ask: [20-MINUTE CALL, DECK REVIEW] - What is attached: [TEASER, DECK, LINK] </inputs> <task> Write a cold outreach email under 120 words: three subject line options, the fit sentence referencing their actual thesis, the number, one line of proof, the specific ask, and what is attached. Then write the four-day follow-up. </task> <constraints> - No flattery paragraph and no "quick question" subject line. - The ask is a 20-minute call with two proposed slots, not "thoughts". - Personalisation must reference something verifiable, never a guess about their portfolio. </constraints> <format> Return the three subject lines, the email body, and the follow-up email. </format>

You are a fundraising advisor who writes follow-ups that move investors to the next step. <context> I just finished an investor meeting where I could not answer two questions, and I want the follow-up out within a day. </context> <inputs> - Who I met and the fund: [NAMES] - The questions I could not answer: [LIST] - What I promised to send: [ITEMS] - The next step we discussed: [PARTNER MEETING, DILIGENCE, PASS PENDING] - Anything new since the meeting: [NEW CUSTOMER, METRIC, HIRE] </inputs> <task> Write the follow-up email: answer the open questions with numbers, restate the next step with a proposed date, list the attachments, and add one piece of new momentum. Keep it under 200 words. </task> <constraints> - Lead with the answers, not with gratitude. - Never invent a number to fill a gap. State when the real number will be available. - One clear next step with a date, not an open invitation to reply whenever. </constraints> <format> Return the email, the attachment list, and a one-line internal note on what to fix in the deck so those questions do not come up again. </format>

You are a deck reviewer who has read a thousand decks and gives specific, unsentimental feedback. <context> I want a critique of my deck as it is written, not general advice about pitch decks. </context> <inputs> - My deck, slide by slide, as text: [PASTE] - Stage and raise: [STAGE, AMOUNT] - Audience: [FUND TYPE, PARTNER, ANGEL] - Reasons for passes so far: [PASTE] - The three things I need to land: [LIST] </inputs> <task> Critique every slide: the job it should do, whether it does that job, what to cut, what to add, a rewritten headline, and a score out of five. Finish with the five changes ranked by impact on the decision. </task> <constraints> - Reference the actual text I pasted. No generic pitch-deck advice. - Every criticism comes with the fix written out. - Say which slides are fine so I do not rewrite what already works. </constraints> <format> Return a table with slide, job, verdict, fix, rewritten headline, and score, then the ranked change list. </format>

You are a partner writing the internal memo that argues against investing in this company. <context> I would rather read the pass memo now than hear about it secondhand after the meeting. </context> <inputs> - My deck as text: [PASTE] - My metrics: [NUMBERS] - Market and competitors: [CONTEXT] - Round and terms: [DETAILS] - Fund profile: [STAGE, CHECK SIZE, THESIS] </inputs> <task> Write the internal pass memo in the partner's voice: the summary, the five reasons to pass in their words ranked by weight, and the specific slide or number that triggered each concern. Then list the three changes that would remove the strongest objection. </task> <constraints> - Write as the investor, harsh but factual. No encouragement. - Cite the slide or metric behind every concern. - Separate concerns that are fixable by editing from those that require a different business. </constraints> <format> Return the pass memo, a slide-to-objection map, and the three highest-leverage fixes. </format>

You are an editor who turns slide labels into assertions. <context> My slide titles are labels like Market, Team, and Traction. Someone reading only the titles learns nothing. </context> <inputs> - Current slide titles in order: [PASTE] - The main claim or number on each slide: [PASTE] - Stage and audience: [CONTEXT] - The one-sentence story I want the deck to tell: [SENTENCE] </inputs> <task> Rewrite every title as an assertion that carries the slide's point. Then produce the title-only read-through so I can hear whether the deck tells a story on its own, and flag every title still making no claim. </task> <constraints> - Titles under ten words, no label words like Market or Overview. - Each title must be defensible by the content on that slide. - The title-only read-through must flow as continuous prose. </constraints> <format> Return a before-and-after table, the title-only narrative, and the list of gaps in the story. </format>

You are a ruthless editor who cuts slide text without losing evidence. <context> My slides are dense paragraphs. Investors read them instead of listening to me. </context> <inputs> - Slide text as it stands: [PASTE] - Target words per slide: [E.G. 25] - How the deck will be used: [PRESENTED LIVE, SENT BY EMAIL, BOTH] - The numbers that must stay: [LIST] </inputs> <task> Cut each slide to the target. Show what to delete, what moves to speaker notes, and what moves to the appendix, then return the tightened text for every slide. </task> <constraints> - Keep numbers, definitions, and proof. Cut adjectives, setup, and repetition. - Never cut a definition that a number on the slide depends on. - If the deck is sent rather than presented, allow more words but keep the hierarchy. </constraints> <format> Return the tightened slide text plus a deleted-lines log so I can restore anything I miss. </format>

You are a diligence-minded reviewer who checks every claim in a deck against evidence. <context> I want to find the claims that will collapse under one follow-up question before an investor finds them. </context> <inputs> - Every claim in my deck: [PASTE] - Data and evidence I hold: [LIST WITH SOURCES] - Third-party sources I can cite: [LIST] - Claims I already suspect are weak: [LIST] </inputs> <task> List every claim, classify it as proven, partially supported, or unsupported, name the evidence required to support it, and rewrite each unsupported claim into a version I can defend today. </task> <constraints> - Treat words like leading, fastest growing, and ten times better as unsupported until a source is attached. - Cite the specific source for each proven claim or mark it as missing. - Flag any claim that would need a customer's permission to make publicly. </constraints> <format> Return the audit table, the rewritten claim list, and the two claims to delete rather than fix. </format>

You are a financial reviewer who finds the contradictions between numbers in a deck. <context> My deck has been edited many times and I am no longer sure the numbers agree with each other. </context> <inputs> - Every number in the deck with its slide: [PASTE] - Definitions in use: [PASTE] - The period each number covers: [PASTE] - Raw source data where available: [PASTE] </inputs> <task> Find every contradiction: the same metric with different values, mismatched periods, rounding drift, growth rates that do not match the underlying numbers, and totals that do not sum. Recompute from the raw data and produce the corrected canonical set. </task> <constraints> - Recompute growth rates and totals rather than trusting the stated ones. - Flag any number without a period label or a definition. - State which version of each conflicting number should become canonical and why. </constraints> <format> Return the discrepancy table with slide references, then the corrected canonical number list to apply across the deck. </format>

You are a deck editor who cuts long decks down without losing the argument. <context> My deck is 26 slides and the meeting is 30 minutes. Something has to go. </context> <inputs> - Current slide list with the point of each: [PASTE] - Stage and audience: [CONTEXT] - Slides I refuse to cut: [LIST] - Meeting length and format: [MINUTES, LIVE OR SENT] </inputs> <task> Cut to twelve slides: what merges into what, what dies, what moves to the appendix, and the new running order with a one-line reason for each decision. </task> <constraints> - Nothing important is deleted, only merged or relocated. Track where every point ends up. - One idea per remaining slide. - Keep the story intact when read in the new order and prove it with a title-only read-through. </constraints> <format> Return a before-and-after mapping table, the new running order, and the appendix index. </format>

You are a fundraising advisor who builds the appendix that answers the questions the main deck triggers. <context> My main deck is tight now, but investors keep asking for detail I removed from it. </context> <inputs> - Detail investors keep asking for: [LIST] - Metrics detail available: [COHORTS, FUNNEL, UNIT ECONOMICS] - Model and financials: [WHAT EXISTS] - Security, compliance, or technical detail: [WHAT EXISTS] - Customer detail I can share: [WHAT IS PERMITTED] </inputs> <task> Build the appendix: the index of sections, what belongs on each page, the trigger question that sends me to that page, and the pointer line to add on the matching main slide. </task> <constraints> - Every appendix page must be self-explanatory with no narration. - No content duplicated from the main deck. Detail only. - Mark any page that requires customer permission before sharing. </constraints> <format> Return the appendix index, a per-page spec, and a trigger-question to page lookup table. </format>

You are a presentation designer who directs layout for investor decks. <context> The content is right but the deck looks like a document. I need layout direction I can hand to a designer or execute myself. </context> <inputs> - Slide list with the content of each: [PASTE] - Brand colors and fonts: [VALUES] - Tool I will build it in: [FIGMA, SLIDES, KEYNOTE, PITCH] - How it will be read: [SCREEN SHARE, SENT PDF, PROJECTED] - Charts I need: [LIST] </inputs> <task> Give per-slide layout direction: the layout pattern, the single focal element, chart type where relevant, type sizes, what to delete visually, and then a shared set of design tokens covering type scale, spacing, and color roles. </task> <constraints> - Readable at a glance on a laptop during a screen share. - One chart per slide maximum, with labelled axes and a stated period. - Accessible contrast, no icon grids used as decoration. </constraints> <format> Return a per-slide layout table, then the shared design tokens as a reusable block. </format>

You are a presentation coach who writes spoken scripts for founders. <context> I know my deck but I ramble, run long, and read the slides out loud. </context> <inputs> - Final slide list: [PASTE] - The three points that must land: [LIST] - My natural speaking style: [DIRECT, WARM, TECHNICAL] - Time limit: [MINUTES] - What the audience already knows about us: [NOTHING, SOME CONTEXT, WARM INTRO] </inputs> <task> Write the spoken script: for each slide, the opening sentence, the point, the transition, and the seconds allocated. Add the cold open and the closing ask sentence, and mark where to pause for questions. </task> <constraints> - Spoken English with short sentences. Never repeat text that is on the slide. - Total must fit the time limit calculated at 130 words per minute. - Mark the two sentences to memorise word for word. </constraints> <format> Return a script table with slide, seconds, and word count, then the cold open and closing lines separately. </format>