Resources · 19 of 20
Prompt Library
A reusable collection of strong, copyable prompts for the everyday tasks of private equity deal work.
When a teaser lands at 6pm and a partner wants a read by morning, you do not want to be composing a prompt from scratch. This is the page to bookmark for exactly those moments. Below is a working collection of prompts for the tasks you do most often, each built on the four-part shape (Role, Task, Format, Scrutiny) from Anatomy of a Good Prompt. They are starting points, not scripts. Paste one in, adjust the specifics to your situation, attach the relevant document, and refine from there.
Each part sits in its own labeled section, such as <task>. The labels help Claude keep the instructions, the format, and the checks apart. They are plain text, not code: type over anything inside them.
Reading and summarizing documents
The bread and butter: turning a long document into something you can act on quickly, without losing the things that matter.
# Version: 1.0 # Tested on: Claude [model name], [month and year], by [your name] <role> You are assisting a private equity associate with an initial screen of Project Atlas, a mid-market B2B field-service software business. The attached information pack is your only source. </role> <task> Produce two things from the attached document, in this order: 1. A one-page screening summary under these headings: - Business overview: what the company does and how it makes money - Financial highlights: revenue, EBITDA, growth, and margins - Investment thesis: why this could be attractive - Key risks and open questions 2. A table of the headline financials for each year shown: revenue, gross margin, EBITDA, EBITDA margin, year-on-year growth, and net debt, with a column citing the page or section each figure came from. </task> <format> Bullets under each summary heading, one page at most. The table follows the summary under its own heading. </format> <scrutiny> Use only what the document states. Where a figure is missing, write "not stated" rather than estimating, because these numbers may be carried into later analysis. After the table, flag any figure that looks inconsistent with the others or rests on an assumption I should check, and list anything the document is vague about that I should press management on. </scrutiny>
Risk and due diligence
Surfacing what could go wrong, and turning a data room into a focused list of things to chase.
# Version: 1.0 # Tested on: Claude [model name], [month and year], by [your name] <role> You are assisting a private equity associate on due diligence for Project Atlas, a mid-market B2B field-service software business. Work from the attached document. </role> <task> 1. Identify the material risks to this investment. Group them under commercial, financial, operational, and legal/regulatory. For each, give the risk in one line and a one-line "so what" explaining why it matters to the investment case. Rank the full list so the most deal-relevant risks come first. 2. Turn those risks into the due-diligence questions I should put to the seller and management, organized by workstream: commercial, financial, technology, people, legal. For each question, note which risk it tests, why it matters, and what a concerning answer would look like. </task> <format> Part 1 as a ranked list grouped by category. Part 2 as one table per workstream with the columns: question, risk it tests, why it matters, concerning answer. Within each workstream, put the questions that could change the thesis first. </format> <scrutiny> Base every risk on the document. If the document hints at a risk without stating it plainly, label it "to confirm" rather than presenting it as fact. Mark any question that exists because the document is silent or unclear on a point, since those gaps are often where diligence finds the real issues. </scrutiny>
Market and competitive work
Framing the market and the competitive picture, kept honest about what is established fact versus general knowledge.
# Version: 1.0 # Tested on: Claude [model name], [month and year], by [your name] <role> You are assisting a private equity associate assessing the market for Project Atlas, a B2B field-service software business serving mid-market customers. </role> <task> 1. Frame the market: the relevant segments, the main demand drivers, and the structural trends that would support or undermine an investment thesis here. 2. Compare Atlas with two or three typical competitors in mid-market field-service software, on: target customer, product breadth, pricing model, apparent strengths, apparent weaknesses, and the main competitive risk to Atlas. 3. Name the three data points I would most want before relying on any market sizing. </task> <format> Part 1 as short headed sections. Part 2 as a table with Atlas and each competitor as columns and the dimensions above as rows. Part 3 as a numbered list. </format> <scrutiny> Label each point as coming from the attached materials or from general sector knowledge, because the two carry different weight in a committee paper. Give no specific market-size figures; say what data would be needed instead. Write "not stated" wherever the materials do not cover a dimension. </scrutiny>
Management and process
Sharpening the questions you put to people, and preparing for the conversations that decide a deal.
# Version: 1.0 # Tested on: Claude [model name], [month and year], by [your name] <role> You are assisting a private equity associate preparing for a management meeting on Project Atlas. Work from the attached materials. </role> <task> Draft the questions I should ask, grouped by theme: growth and pipeline, unit economics, customer retention, team and key-person risk, and the plan for the next three years. For each question, add the one-line reason it matters and what answer would reassure me versus concern me. </task> <format> One short list per theme, with the must-ask questions first in each. </format> <scrutiny> Flag any question prompted by a gap or inconsistency in the materials, since those are the ones management will be least prepared for. </scrutiny>
Writing and drafting
Turning analysis into the documents that carry it forward, in the right register.
# Version: 1.0 # Tested on: Claude [model name], [month and year], by [your name] <role> You are assisting a private equity associate drafting an investment committee paper on Project Atlas. Work only from the attached analysis. </role> <task> Write two pieces for the paper: 1. An executive summary for a busy partner: no more than 200 words, in three short paragraphs covering what the business is, why it could be attractive, and the main concern. 2. The investment thesis section: about 300 words, structured as the thesis, the value-creation angle, and the main risk to the thesis. </task> <format> The executive summary first, then the thesis section, each under its own heading. Write in a measured, specific register suited to an investment committee: plain and confident, with no marketing language or superlatives. Close with a short "Open questions" list. </format> <scrutiny> Base every claim on the attached analysis. Where something important is missing, add it to "Open questions" instead of filling the gap, because the committee will read every sentence as supported. </scrutiny>
Reshaping and refining
The smaller moves that save time once you already have a draft.
# Version: 1.0 # Tested on: Claude [model name], [month and year], by [your name] <role> You are helping a private equity associate refine material for an investment committee audience. </role> <task> The material below is either one draft or two versions or options, labeled A and B. If it is one draft: rewrite it for an investment committee, measured, precise, and free of marketing language, keeping every substantive point intact. If it is A and B: compare them, then give your one-line read on the trade-off and say which dimensions are most decision-relevant. If A and B are two drafts of the same text, finish by rewriting the stronger one as above. </task> <format> For a comparison, a table with the dimensions that matter as rows and A and B as columns, then your read. Any rewrite follows under its own heading. </format> <scrutiny> Add no new claims or figures. If a sentence makes a claim the underlying material would not support, flag it rather than rewriting around it. Where you are inferring rather than working from stated facts, label it. </scrutiny> <material> [paste one draft, or label two versions or options "A:" and "B:"] </material>
How to use these well
Four habits make the difference between a prompt that helps and one that hinders.
- Adapt the specifics. Swap Project Atlas for your situation, and make the role line match the real task.
- Always keep the scrutiny section. The “not stated” and “flag it” lines are what keep the output honest. Do not delete them to make a prompt shorter.
- Fill in the header once you have tested it. The
Tested online records which model a prompt was checked on. Run it on a real job, then add the model, the date, and your name, as From Prompt to Template explains. - Refine in place. A first answer that is 80 percent right is normal. Correct the rest with a follow-up rather than starting over, as Anatomy of a Good Prompt describes.
A single prompt gets you a fast, useful answer. A playbook strings several of them into a full workflow, from first screen to committee paper. When you are ready to go end to end, see CIM Analysis, Due Diligence, Market Research, Financial Review, and Investment Memo. If a term in any prompt is unfamiliar, the Glossary has it in plain English.