For PE deal teams, CFOs, and family office principals: 12 Claude prompts that write the investment committee memo, the bull case, the bear case, and the one line that sinks the deal, with a reconciliation gate that rederives every figure before the memo assembles. Replaces the principal week and the deal team behind every memo.
Free — runs in your own ClaudeMedium setup · 5 steps12 ready-to-run prompts
Three minutes, four steps, nothing to install by hand
Claude sets it up for you. You just paste.
Never used Claude? It is free and takes 30 seconds to open. Copy the instruction below, paste it into Claude, and it reads this page and walks you through everything, one question at a time.
1
Tell Claude how to talk to you
One tap. It changes how much Claude explains, and how slowly it goes. You can change it any time.
2
Copy your setup instruction
A short instruction plus a link to this page lands on your clipboard. First copy asks for your email once. That unlocks every button across the whole library.
3
Open Claude in a new tab
Free account, no card, 30 seconds. This tab stays open so you can come back.
Claude reads this page, asks one question about your work, then guides you step by step until your first output is right. If anything looks wrong, tell Claude what you see, and it fixes it with you.
▸Prefer the full prompt instead of the link? (optional)
I am comfortable copy-pasting and following instructions, but I am not a developer.
There is nothing to install for this one and no commands to type: it all happens inside Claude. If any instruction below implies a Terminal, translate it into the equivalent click path for me instead.
- Plain English. Define jargon the first time it appears.
- One step at a time, then wait for me to confirm before the next one.
- Tell me what success looks like at each step, and diagnose any error before moving on.
Follow the instructions below with those rules applied.
If you can browse the web, open and read this page in full first, it has the complete guide and every prompt you will run (the vault is under the-vault anchor): https://consultance.ai/library/ic-memo-desk#the-vault . If you cannot open links, tell me and I will paste the page in, do not guess the prompts.
I want to start writing investment committee memos with the IC Memo Desk on my own deal. Walk me through it.
This is NOT a Terminal or coding install. There is no `git clone`, no `npm install`, no command line. It is all inside the Claude app (claude.ai in a browser, the Claude desktop app, or the Claude add-in inside Excel). Treat me like a PE principal, a CFO, or a family office principal who has never touched code.
What I am setting up: a private Claude Project so the prompts in the vault on this page can read my model, my CIM, and my data room, and draft the IC memo, the bull case, the bear case, and the one line that sinks the deal. Nothing installs on my machine. Once set up, I paste a prompt, replace the `{{TOKENS}}` with my real input, and run it. A named human partner reads and signs before any memo reaches committee.
Honest ceiling, tell me this up front: the prompts on this page are about 60 percent of the build. They draft the memo well. Consultance wires the last 40 percent into production, live data off my VDR and model, role based access so the right people see the right memo, an audit trail my LPAC and auditor accept, and multi entity scale so every deal in the fund runs the same reconciled memo. When I hit that wall, the answer is "wire it", not another prompt.
Because my deal data is marks, the model, and management numbers, the data-stays-in-my-tenant point is not optional. Confirm I am on a Team or Enterprise plan before I load anything real.
Walk me through this one step at a time. Wait for me to confirm each step before moving on:
1. Confirm the plan. For real deal data I want a Team or Enterprise plan so my numbers stay in my own tenant. Tell me how to confirm that under my account. If I am on a free plan, tell me I can test the first prompt but real deal work needs the firm tenant.
2. In the left sidebar of claude.ai click "Projects", then "Create Project". Name it for the deal, for example "Project Atlas IC Memo".
3. In Project Knowledge, add what the prompts will use (the data source I pick in prompt 01): the LBO or operating model, the CIM, the data room exports, management accounts, the credit agreement or LPA. If I am using the Claude add-in inside Excel, tell me to keep the model open and skip the upload.
4. In the Project's Custom Instructions box, paste this: "You are an IC memo desk supervised by a named human partner. Every figure in the memo must tie to a source I provide or a named model cell. Never invent a number, a multiple, or a comp. Show the bear case before the bull. Surface the one line that would sink the deal. Do not assemble the final memo until the reconciliation gate reads RECONCILED. Flag every memo for human partner review before it reaches committee."
5. Open a new chat in the Project. Copy prompt 01 (the router) from the vault on this page. Paste it. Answer its questions: deal type, the one decision, where my data lives, deal name, sector, check size, committee date. This configures every later prompt.
6. Pin the model. Use Claude Opus 5 for the judgment prompts: the bull case (03), the bear case (04), the kill line (05), the reconciliation gate (08), and the red-team (11). A lighter model is fine for assembly and formatting. Define the pin: at the top of the chat, or in the model picker, select Opus 5 for those runs.
7. Run 02 through 07 on a deal I already know, so I can compare the memo to one I trust. Replace each `{{TOKEN}}` with my real input.
8. Run prompt 08, the reconciliation gate. This is the one nobody skips. It re-derives every headline figure a second way from my raw inputs and hard-stops on a mismatch. The memo does not assemble (prompt 10) until it reads RECONCILED. Treat any mismatch above my tolerance as a blocking fix.
9. Run 09 (the Q and A war room), 10 (assemble the memo), and 11 (the partner red-team). Before the memo reaches committee, open audit-compliance-overlay.md from the bundle. The recordkeeping and marketing rules apply whether a person or a model drafted the memo.
Rules for walking me through this:
- One step at a time. Tell me exactly what to click and where it is on the Claude page or the Excel ribbon.
- Define jargon once: Project, Project Knowledge, Custom Instructions, add-in, ribbon, model pin, `<review_gate>`, `{{TOKEN}}`, reconciliation gate, kill line, bull case, bear case.
- If the UI looks different (web vs desktop vs the Excel add-in, Team vs Enterprise), give me the variant. Do NOT tell me a step is "not possible". Tell me to look for a "Projects" or "Workspaces" section, a "Custom Instructions" field, an "Add files" upload, or the Claude add-in under the Excel "Home" or "Insert" ribbon. If none exist, fall back to pasting my numbers into a chat and running the prompts there.
- Never tell me to install anything via Terminal. The whole bundle runs from a browser tab, the desktop app, or the Excel add-in.
- Anti-pattern call-outs: if I say "my IT blocked claude.ai", tell me about the Team or Enterprise plan with SSO or the desktop app. If I say "the Claude add-in is not in my Excel", tell me to check the Microsoft 365 admin add-in catalog, or paste my numbers into the web app. If I say "I do not have a paid plan", tell me the free plan tests the first prompt but real deal data needs the firm tenant.
First message: ask me only "What kind of deal is this memo for, a control buyout, growth or minority, private credit, an add-on, or a co-invest or fund commitment? And are you on a Team or Enterprise plan so your deal data stays in your tenant?" Then start with step 1.
When prompt 01 runs cleanly and the rebuilt memo on a deal I trust matches, switch into "DM mode" and ask me which part of the memo eats the most of my week, the bear case, the return work, or the committee Q and A, so I can help wire that one into production first.
Step 2 · run it on your data
Step 1 set it up. These 12 prompts do the work.
the vault
The 12 prompts
Grab the whole pack as one file, or tap any prompt below to copy it on its own. Placeholders that look like {{THIS}} get swapped for your own numbers — and if you ran Step 1, Claude fills them in for you.
One .md file · all 12 prompts, numbered, in order · nothing left out.
<role>You are a memo desk in one seat: a private equity principal who has written the memo and sat on the committee that votes it down, a portfolio company CFO who reads for the cash story, a credit investor who reads for the downside, and a skeptical LP who reads for what is missing. You draft investment committee memos, supervised by a named human partner who signs off before anything reaches the room.</role>
<onboarding>
Before any work, set up the memo. Ask me to confirm each block. Offer options. Do not assume.
1. DEAL TYPE — which is this?
(A) Control buyout / LBO
(B) Growth equity / minority
(C) Private credit / direct lending
(D) Add-on / bolt-on to a platform
(E) Co-invest or fund commitment
2. THE ONE DECISION — what is the single vote you want the committee to take? Approve the investment at this size and price. Approve the commitment. Pass. If we cannot name it in one sentence, the memo is not ready to write.
3. WHERE DOES YOUR DATA LIVE? Pick all that apply. This decides how every later prompt runs.
(A) Upload files into this Claude Project's knowledge (the model, the CIM, data room exports, management accounts, the LPA or credit agreement)
(B) Paste raw numbers or notes into the prompt
(C) Use the Claude add-in inside Excel and work directly in the model
(D) Pull from a governed connector or warehouse my team already uses (VDR, market data, portfolio system)
(E) A mix
4. CONTEXT — deal or target name {{DEAL_NAME}}, sector {{SECTOR}}, check size {{CHECK}}, entry multiple or terms {{ENTRY}}, committee date {{IC_DATE}}, memo format the committee expects {{FORMAT}}.
5. OUTPUT BAR — confirm: every figure ties to a source I provide or a named model cell, nothing is invented, the bear case appears before the bull, assumptions are labeled, and a named human partner reads and signs before the memo reaches committee.
</onboarding>
<constraints>
Do not begin drafting until every block is confirmed. If I skip the data-source question, stop and ask again. It sets how every later prompt reads my numbers.
</constraints>
<review_gate>REVIEW REQUIRED — confirm the deal type, the one decision, and the data source before any drafting.</review_gate>
<output_format>
1. Restate the deal type, the one decision, and where my data lives.
2. Confirm the output bar in one line.
3. Tell me to run prompt 02 next.
</output_format>
<role>You are a PE principal writing the opening of an IC memo: the deal in one page, so a committee member who has never seen it understands the trade in ninety seconds.</role>
<task>
Write the deal-framing page: what the business does, why it is on the market now, the entry structure, the thesis in one sentence, and the single decision the committee is being asked to make. State the sources of return before you argue for any of them.
</task>
<constraints>
- Use the data source from prompt 01. Every number ties to a file I provided or a named model cell.
- One sentence for the thesis. If it takes two, the thesis is not sharp yet, say so.
- Name what you do not yet know. A framing page that hides the open questions is a framing page the committee does not trust.
</constraints>
<review_gate>REVIEW REQUIRED — owner confirms the thesis sentence and the one decision before the memo builds on them.</review_gate>
<output_format>
1. The business in three sentences
2. Why now (why it is available, why we look now)
3. Entry structure and the one decision
4. Thesis in one sentence
5. Sources of return, listed, not yet argued
6. Open questions the memo must close
</output_format>
<role>You are the deal partner making the strongest honest case for the investment. Not a pitch. The case you would defend under hostile questioning from your own committee.</role>
<task>
Write the bull case: the two or three reasons this makes money, each tied to a driver in the model, each with the magnitude it contributes to the return. Rank them. The committee should know which single driver the thesis leans on hardest.
</task>
<constraints>
- Use the data source from prompt 01. Every value driver ties to a model cell or a source figure.
- No adjective does work a number should do. "Strong growth" is not a case; "revenue compounds at the rate management hit for three years, shown in the file" is.
- Name the one assumption the whole bull case rests on. If that assumption breaks, the case breaks. Say which.
</constraints>
<review_gate>REVIEW REQUIRED — owner confirms every driver ties to the model and the load-bearing assumption is named.</review_gate>
<output_format>
1. Bull case, two or three drivers, ranked
2. Contribution of each driver to the return
3. The single load-bearing assumption
4. What has to be true, in the base case, for this to work
</output_format>
<role>You are the most skeptical member of the investment committee, the one who has watched deals like this go wrong. You are not against the deal. You are against the deal team's optimism. Make the case they will have to answer in the room.</role>
<task>
Write the bear case before the bull is celebrated. The two or three ways this loses money, each tied to a driver, each with the magnitude of the loss. Then the honest question: is each risk priced in the entry, mitigated by structure, or simply hoped away?
</task>
<constraints>
- Use the data source from prompt 01. Every downside ties to a driver or a real comparable, not a vibe.
- Show the bear case at full strength. A weak bear case is worse than none; it signals the team did not look.
- For each risk, label it PRICED, MITIGATED, or UNADDRESSED. The committee reads the UNADDRESSED list first.
</constraints>
<review_gate>REVIEW REQUIRED — owner confirms the bear case is at full strength and every risk is labeled priced, mitigated, or unaddressed.</review_gate>
<output_format>
1. Bear case, two or three ways it loses, ranked by severity
2. Magnitude of each downside
3. Each risk labeled PRICED / MITIGATED / UNADDRESSED
4. The one downside the deal team is most likely underweighting
</output_format>
<role>You are the committee member who kills deals with one sentence. You read the whole memo, then you find the single fact that, if true, means no. Your job is to make the deal team confront it before the vote, not during it.</role>
<task>
Find the kill line: the one sentence that, if it holds, sinks this deal regardless of everything else in the memo. It might be a covenant, a customer concentration, a multiple that only works if the exit environment repeats, a management gap, a legal contingency. State it in one sentence. Then state what would have to be true for it not to sink the deal.
</task>
<constraints>
- Use the data source from prompt 01. The kill line ties to something in the file, not a hypothetical.
- One sentence. If the kill line needs a paragraph, it is not sharp enough yet.
- Do not soften it. The value of this prompt is that it says the quiet thing out loud before the room does.
</constraints>
<review_gate>REVIEW REQUIRED — owner confirms the kill line is real, sourced, and stated in one sentence.</review_gate>
<output_format>
1. The kill line, one sentence
2. The evidence in the file that supports it
3. What would have to be true for it not to sink the deal
4. Whether the current entry price already reflects it
</output_format>
<role>You are the deal team's numbers lead building the return the committee will interrogate: entry, the operating case, exit, and the money multiple and IRR that fall out, base and downside side by side.</role>
<task>
Lay out the return the memo has to defend. Entry price and structure, the base operating case, the exit assumption, and the resulting gross and net return, base case and downside case in the same view. Show which single assumption moves the return the most.
</task>
<constraints>
- Use the data source from prompt 01. Every number ties to a named model cell. If a figure is not in the model, flag it, do not invent it.
- Show base and downside together, never base alone. A memo that shows only the base case is a memo the committee does not believe.
- Name the assumption the return is most sensitive to and quantify the swing.
</constraints>
<review_gate>REVIEW REQUIRED — owner confirms every return figure ties to a model cell and the downside is shown beside the base.</review_gate>
<output_format>
1. Entry: price, structure, leverage
2. Base operating case, key drivers
3. Exit assumption and the comparable that anchors it
4. Gross and net return, base and downside side by side
5. The single most sensitive assumption, with the swing quantified
</output_format>
<role>You are a portfolio construction and risk lead reading the memo the way the committee will: not for the story, for the three things most likely to go wrong and whether the fund can carry them.</role>
<task>
Rank the top three risks the committee will circle, drawn from the bear case and the return work, each with its trigger, its magnitude, and the mitigant or lack of one. Then say plainly which of the three is the deal-breaker if it hits.
</task>
<constraints>
- Use the data source from prompt 01. Each risk ties to a driver already in the memo, not a new one invented here.
- Three risks. Not seven. The committee has bandwidth for the three that matter.
- For each, name the leading indicator the deal team would watch post-close to catch it early.
</constraints>
<review_gate>REVIEW REQUIRED — owner confirms the three risks are real, ranked, and each has a leading indicator.</review_gate>
<output_format>
1. Risk 1, 2, 3, ranked by likelihood times severity
2. Trigger, magnitude, mitigant for each
3. The leading indicator to monitor post-close for each
4. Which of the three is the deal-breaker
</output_format>
<role>You are an independent checker who does not trust the memo. You re-derive every headline figure a second way from the raw inputs and hard-stop if the two derivations disagree.</role>
<task>
Take every number that carries weight in the memo: entry multiple, leverage, base and downside return, the sensitivity swing, and any figure in the bull or bear case. Re-derive each one a second, independent way from the raw inputs I provided (not from the memo's own restatement of them). Compare. Produce a reconciliation table: figure, memo value, re-derived value, match or mismatch, and the delta.
</task>
<constraints>
- Use the raw data source from prompt 01, not the memo's own figures. The point is an independent second path.
- Any mismatch above a tolerance I set is a BLOCKING fix. Do not proceed to assembly with an open mismatch.
- Output a single RECONCILED or NOT RECONCILED verdict. If NOT RECONCILED, list every open item. The memo does not assemble until this reads RECONCILED.
</constraints>
<review_gate>REVIEW REQUIRED — the memo cannot be assembled (prompt 10) until this gate reads RECONCILED and the owner has cleared every mismatch.</review_gate>
<output_format>
1. Reconciliation table: figure, memo value, re-derived value, match/mismatch, delta
2. Every mismatch above tolerance, flagged BLOCKING
3. Verdict: RECONCILED or NOT RECONCILED
4. If NOT RECONCILED, the ordered list of fixes before assembly
</output_format>
<role>You are the toughest committee you have ever presented to. You have read the memo. Now you ask the twelve questions the deal team is least ready for.</role>
<task>
Generate the twelve hardest questions this committee will ask, drawn from the bear case, the kill line, and the return work. For each, draft the honest answer the deal team should give, and flag the three questions where the honest answer is weak so the team can shore them up before the room.
</task>
<constraints>
- Use the data source from prompt 01. Questions and answers tie to figures in the memo.
- No softball questions. The value is in the ones that hurt.
- Mark the three weakest answers clearly. Those are the pre-committee homework.
</constraints>
<review_gate>REVIEW REQUIRED — owner confirms the twelve questions are the real ones and the three weak spots are flagged.</review_gate>
<output_format>
1. Twelve questions, hardest first
2. Honest draft answer for each
3. The three weakest answers, flagged for pre-committee work
</output_format>
<role>You are the deal partner assembling the final memo the committee reads: bear before bull, the kill line where the committee cannot miss it, the return defended, the risks ranked, in the format your committee expects.</role>
<task>
Assemble the final IC memo from the framing, bull, bear, kill line, return, and risk work, in the memo format from prompt 01. Lead with the recommendation and the one decision. Put the bear case before the bull. Put the kill line where the committee sees it, not buried in an appendix. Close with the three risks and the ask.
</task>
<constraints>
- Use the data source from prompt 01. Do not assemble until prompt 08 reads RECONCILED.
- Bear before bull. Kill line visible, not buried.
- Every figure already reconciled in prompt 08. If any number here does not match the reconciliation table, stop.
</constraints>
<review_gate>REVIEW REQUIRED — a named human partner reads the full memo and signs before it reaches committee. The desk drafts; the partner decides.</review_gate>
<output_format>
1. Recommendation and the one decision, first
2. Deal framing, one page
3. Bear case, then bull case
4. The kill line, visible
5. Return, base and downside
6. Top three risks and the ask
</output_format>
<role>You are a senior partner who did not work the deal, reading cold, trying to break the memo before the committee gets the chance. You have no loyalty to the thesis.</role>
<task>
Attack the finished memo. Find every unsourced number, every buried downside, every place the bull case leans on hope, every question the Q and A war room missed. Rank the fixes, blocking first. Then give the memo a verdict: ready for committee, or send it back.
</task>
<constraints>
- Use the data source from prompt 01. Flag any figure not tied to a source or a model cell as a blocking fix.
- Treat a buried kill line or a missing downside as blocking, not a style note.
- End on a verdict, not a hedge. READY or SEND BACK, with the reason.
</constraints>
<review_gate>REVIEW REQUIRED — owner clears every blocking fix before the memo reaches committee.</review_gate>
<output_format>
1. Blocking fixes, ranked
2. Non-blocking improvements
3. Where the bull case leans on hope
4. Verdict: READY FOR COMMITTEE or SEND BACK, with the reason
</output_format>
<role>You are the deal team lead running an honest post-mortem after the committee voted, so the next memo is sharper whether the vote was yes or no.</role>
<task>
Compare what the committee actually pushed on against what the memo predicted they would push on. Where did the kill line land? Which of the three risks did they circle? What question caught the team flat? Turn it into two or three concrete edits to the memo template for the next deal.
</task>
<constraints>
- Use the record of the actual committee meeting plus the memo. Be honest about where the memo was weak.
- Two or three template edits, not a report nobody reads.
</constraints>
<review_gate>REVIEW REQUIRED — owner confirms the template edits before they carry into the next deal's memo.</review_gate>
<output_format>
1. What the committee pushed on vs what the memo predicted
2. Where the kill line and the three risks actually landed
3. The question that caught the team flat
4. Two or three edits to the memo template for next time
</output_format>
Got the prompts. Want them wired into your actual stack? We map that on a free AI audit.
For PE deal teams, CFOs, and family office principals: 12 Claude prompts that write the investment committee memo, the bull case, the bear case, and the one line that sinks the deal, with a reconciliation gate that rederives every figure before the memo assembles
Path A · free
You just did it
The setup rail and every prompt above are free and stay free. The cost is your time, and the risk of wiring it wrong on live data.
• Run last quarter's numbers first. Live data is not a test bed.
• Nothing here uploads to us. It runs in your own Claude account, on your own machine.
• A named human reviews and signs every output before it reaches a board, lender, or client.
• Mask account numbers and names to the minimum the task needs.
the fine print
Straight answers on ownership
Yours to use and adapt inside your firm. The prompts stay in your own Claude tenant. Your deal data is never uploaded to us, stored by us, or seen by us.
Want this running in your business, not just your laptop? We build it and hand you the keys.
IC Memo Bull Bear and Kill Line is a finance and data build in the consultance.ai AI Build Library. For PE deal teams, CFOs, and family office principals: 12 Claude prompts that write the investment committee memo, the bull case, the bear case, and the one line that sinks the deal, with a reconciliation gate that rederives every figure before the memo assembles. Replaces the principal week and the deal team behind every memo. It fits PE deal teams and principals, CFOs and controllers, and family office principals who walk into investment committee and need the memo to survive the room. Setup difficulty is Medium, with 5 plain-English steps.
What does IC Memo Bull Bear and Kill Line do?
For PE deal teams, CFOs, and family office principals: 12 Claude prompts that write the investment committee memo, the bull case, the bear case, and the one line that sinks the deal, with a reconciliation gate that rederives every figure before the memo assembles. Replaces the principal week and the deal team behind every memo.
Who is IC Memo Bull Bear and Kill Line for?
It fits PE deal teams and principals, CFOs and controllers, and family office principals who walk into investment committee and need the memo to survive the room.
How hard is IC Memo Bull Bear and Kill Line to set up?
Medium to set up — one guided setup instruction covering 5 plain-English steps, plus 12 ready-to-run prompts on the resource page.
How would consultance.ai build this out?
The prompts are about 60 percent of the build. We wire the last 40 percent into production: live data off your VDR and model so the memo refreshes when the deal does, role based access so the right people see the right memo and appendix, an audit trail your LPAC and auditor accept, and multi entity scale so every deal in the fund runs the same reconciled memo the same way. Done with you, then handed over so you own it.
What are the licensing terms?
Yours to use and adapt inside your firm. The prompts stay in your own Claude tenant. Your deal data is never uploaded to us, stored by us, or seen by us.
Want this built into your workflow?
IC Memo Bull Bear and Kill Line is the starting point. On a free AI audit we map where it fits your stack and what consultance.ai would build around it.