Explore the build library
consultance.ai
Deal Copilot
Build libraryControlsAboutCase studiesBlogBook an audit
consultance.ai

Systems

WorkflowsDeal CopilotPrivate knowledgeFinance automationBuild library

Company

About usCase studiesBlogContact

Start a conversation

Free AI auditfounder@consultance.ai

Teams

Investment bankingCFOs & financePrivate equityPrivate creditReal estateProperty managementProfessional servicesOther industriesAll industries

Legal

Privacy policyTerms of service

Our approach

Audit the workflow.Build in your environment.Hand over a system you own.
Build something useful © 2026 consultance.ai
Build library/Outside-In Diligence Review

Finance and data

Outside-In Diligence Review

For PE deal teams and corp dev: test every CIM claim against pricing history, hiring, reviews, demand and customer logos, each figure tied to its public source.

Medium setup4 steps13 prompts
Open the setup guide Discuss an implementation
Finance and data

Start with the guided setup

Follow the setup below in your own workspace. For a demonstration of the prompt-copying process, open the optional walkthrough.

Watch the setup walkthrough
Step 1 · setup
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.

    Open claude.ai
  4. Paste, send, and answer one question

    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)
Click to copy
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/outside-in-diligence#the-vault . If you cannot open links, tell me and I will paste the page in, do not guess the prompts.

You are going to help me run the Outside-In Diligence Review on one target, in my own Claude, with my own accounts. Go one step at a time and wait for me after each step.

First question, and only this one: where does the deal material live? A CIM or teaser I can upload here, a folder on my computer, or only the target's website for now?

If I say upload or website only: this is a Claude app job. Tell me to make a private Project, select Claude Opus 5.5 in the model picker, and turn on web search. Do not give me a Terminal install.
If I say a folder: this is a Claude Code job. Tell me to open Claude Code in that folder, one step at a time, and tell me what success looks like after each step.

Then run the first session drill with me:
1. Paste prompt 01, the shared rules. Ask me my thesis and what worries me, then the target, the claims, the competitors, the period and the named sign-off person, one at a time.
2. Run prompt 02 on the fake sample and compare every line to the answer key. If one line does not match, stop and tell me why before we touch real data.
3. Then run the first unit, frame-the-claims, on my real target.
Step 2 · run it on your data

Step 1 set it up. These 13 prompts do the work.

the vault

The 13 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 13 prompts, numbered, in order · nothing left out.
Click to copy
OUTSIDE-IN DILIGENCE REVIEW: SHARED RULES
Every unit in this pack follows these rules. They are written here once.

1. WHERE TO RUN IT
- One target: the Claude app, in a private Project, with web search turned on. Chat is fine here.
- Several targets, or saved web captures and exports in a folder: Claude Code, opened in that folder.
- Deal material that must stay on a managed machine: Claude Code on that machine.
- The same refresh every month while the deal runs: Claude Code, so each run lands as dated files you can compare.
Model: select Claude Opus 5.5 in the model picker. Use Claude Sonnet 5 only for the first bulk read of a large export of reviews or postings.
Privacy: your materials go only to your own Claude account, never to us. Use Team or Enterprise, or turn model training off, before you add the CIM or anything under NDA.

2. BEFORE THE FIRST RUN
First ask the human: what is your thesis on this company, and what worries you most? Record both answers word for word. Every later unit tests the claims that bear on them first.
Then ask for these, one at a time: {{TARGET}} (legal name and website), {{CIM_CLAIMS}} (the CIM or teaser, or the claims typed out), {{COMPETITORS}} (2 or 3 names), {{PERIOD}} (default: the last 24 months), and the named person who signs off.
A missing item gets a plain default, labelled ASSUMED. No CIM means you test the claims on the target's own website and press releases, labelled MARKETING.

3. STOP RATHER THAN GUESS
- A file path or folder listing pasted into chat: this is a Claude Code job. Say so and stop.
- A page you cannot open, or one behind a login: name it, mark it NOT REACHED, and do not score it. Never log in, never use a paid panel the human has not connected, never get round a paywall.
- A figure needs a document from the data room you do not have: ask once, name the document, mark it OPEN.
A stop blocks only the outputs that need the missing input. Everything else still runs.

4. LABELS ON EVERY FIGURE
REPORTED (source, URL or page, and the date you read it) · CALCULATED (show the arithmetic) · ASSUMED (who set it) · MARKETING (a claim in the CIM, deck, website or press release: it makes a question, never a number) · NOT IN SOURCE.
Third party estimates (traffic panels, app rank trackers) are REPORTED with the word "estimate".
A MARKETING figure may feed a calculation only as a conditional: "on the seller's own figures". The result is a test of the claim, never a fact. An outside figure never replaces a data room figure. When they disagree, the gap becomes a question for management.

5. THREE FLAG TAGS
- CHANGES THE DECISION: reaches the final scorecard and the memo section.
- WORTH A QUESTION: only if the answer could move growth, retention or margin by more than 5 points, or the price by more than 5%. Smaller items go on the management question list, untagged.
A claim gets a verdict (SUPPORTED, CONTRADICTED, NOT TESTABLE FROM OUTSIDE, OPEN). A finding that cuts against a claim also gets one flag tag.
- EXPLAINED BY CONTEXT: one line under "Checked, normal". Once closed, it stays closed.
Test every flag against the list below before you raise it. If the "when it is not" test needs data you do not have, close the flag and note the test as not run.

6. LOOKS WRONG, IS NORMAL
- Job postings fall in December, or right after a raise closes | hiring runs in waves | not normal if sales and customer success roles vanish while the plan says growth.
- B2B web traffic estimates are low and jumpy | panels undercount office and VPN visits | read direction over 12 months, never the level.
- A spike in traffic | one launch, press hit or campaign | not normal if the CIM uses that month as the run rate.
- Seasonal demand | payroll and tax peak in Q1, retail in Q4 | compare the same month year on year.
- Reviews skew negative and cluster after a price rise or a redesign | unhappy users write more | under 30 reviews in a period says nothing on its own.
- Employee rating dips after a reorg or layoff, then recovers | one time event | not normal if it stays down in sales or support.
- A customer logo disappears from the website | brand refresh, or the customer asked | not churn unless a second source says so.
- The pricing page becomes "contact sales" | a move upmarket | still check what list prices did before it went dark.
- App rank falls while revenue grows | users moved to web or to enterprise seats | not normal if the CIM says growth is self serve.

7. STANCE
A neutral colleague who has done outside-in diligence many times. Do not hunt for problems. Do not take a claim as proof. Outside data shows direction and raises questions; it does not prove a number.
Use public pages only and respect each site's terms. Do not collect personal data on named employees. Never contact the target's staff, customers or suppliers: the deal is confidential, and outreach can breach the NDA.

8. HANDOFF (every unit ends with this)
Read (and not read) · Outputs · Checks: pass, fail or not run · Open items, with the document needed · Next unit · Sign-off needed from the named person.
Click to copy
Run the pack on this sample before you use real data. Compare your result to the answer key. If one figure does not match, stop and find out why first. One document is missing on purpose, so you see the pack stop instead of guess.

SAMPLE (fake companies; every figure cites its page or evidence row)

CASE A: Harbor Ledger (accounts payable software, annual contracts)
- CIM p.12: "ARR grew 40% to $14.0M, driven by new logos."
- CIM p.14: "Gross revenue retention of 94%."
- CIM p.22: "Contracts renew annually at the current list price."
- Evidence row 1: archived pricing page, captured Jan 2025: Growth plan $400 per month.
- Evidence row 2: archived pricing page, captured Jan 2026: Growth plan $500 per month.
- Evidence row 3: careers page postings for sales roles: 3 open in Jan 2025, 9 open in Jan 2026.

CASE B: Pinecrest Payroll (payroll software for small firms)
- CIM p.8: "Demand is at record highs."
- CIM p.9: "Net revenue retention of 108%."
- CIM p.10: "No price increase in three years."
- Evidence row 4: web traffic estimate: Jan 2026 is 3.0 times Sep 2025. Jan 2026 is 1.18 times Jan 2025. April 2026 is back near the Sep 2025 level.
- Evidence row 5: employee review rating: 4.1 in Feb 2026, 3.4 in Mar 2026, 4.0 in Aug 2026. Press release p.1, 2 Mar 2026: a reorganisation of the product team.
- Evidence row 6: open job postings: 12 in Nov 2025, 4 in Dec 2025, 11 in Jan 2026.
- Evidence row 7: archived pricing page: $6 per employee per month in captures from Jan 2024, Jan 2025 and Jan 2026.
- The customer cohort file from the data room is NOT provided.

ANSWER KEY
| Check | Expected | Working |
| A. Price change | +25% list price | CALCULATED: 500 / 400 - 1 = 25% (rows 1, 2) |
| A. Prior ARR | $10.0M | CALCULATED: 14.0 / 1.40 (CIM p.12) |
| A. Growth split | Price $2.35M, new logos $2.25M, churn -$0.6M, total $4.0M | CALCULATED: retained base 10.0 x 0.94 = 9.4. Repriced 9.4 x 1.25 = 11.75, so price adds 2.35. Churn 10.0 - 9.4 = 0.6. New logos 14.0 - 11.75 = 2.25 |
| A. "Driven by new logos" | CONTRADICTED, CHANGES THE DECISION | New logos are $2.25M of $4.0M. Price is the larger part. New logo growth is 2.25 / 10.0 = 22.5%, not 40% |
| A. Sales hiring | Intent only, WORTH A QUESTION | 3 to 9 postings is intent to hire, not headcount. Ask management for sales headcount by month |
| B. Traffic | EXPLAINED BY CONTEXT | Payroll peaks in Q1. Same month year on year is +18% (row 4): supports demand growth. "Record highs" stays MARKETING |
| B. Employee rating | EXPLAINED BY CONTEXT | One time reorg (press release p.1), recovered to 4.0 by Aug (row 5) |
| B. December postings | EXPLAINED BY CONTEXT | Seasonal dip, back to 11 in Jan (row 6) |
| B. Pricing | SUPPORTED | Same $6 in three captures (row 7) |
| B. Retention 108% | OPEN | Needs the customer cohort file. Outside data cannot test it. Every other check still runs |
| Verdicts | A: reprice the growth case, ask about the price rise. B: no flag changes the decision, one open item |

The trap is Case A. The plausible answer is "40% growth from new logos, supported by sales hiring tripling". It is wrong because the list price rose 25% and every renewal moved to it.
A run that flags any Case B item, or guesses the retention figure, has failed.
Click to copy
<unit name="tools-setup">
Follow the shared rules in prompt 01.
What it does: suggests tools for the jobs in this pack, then installs only the ones the human approves, on the human's own accounts.
Use when: after prompt 02 passes, before the first real run. Skip it if web search and upload are enough.
Needs: the "Needs" lines of every unit in this pack, and the systems the human already uses: web search, the Internet Archive's Wayback Machine (web.archive.org), any traffic, app or review data their firm already licenses, their VDR, and their CRM.
Method:
1. See what is already connected. In Claude Code, run `claude mcp list`. In the Claude app, look up the current steps in Anthropic's own help pages with your web search, then ask the human what their Claude already shows as connected.
2. For each need, search the web yourself for an MCP server or connector published by the vendor the human already uses, or listed by Anthropic. Prefer: official publisher, read only access, sign-in by OAuth, updated in the last 6 months.
3. Show one table: need · tool · publisher · official page URL · read only or read and write · how far it reaches · what it costs on the human's own plan (write "not stated" if the page does not say) · last update (write "no date shown" if none).
4. For each tool the human approves, open its official page and follow its own steps. A server with a command or URL: copy the command word for word, never from memory. In Claude Code the shape is `claude mcp add --transport http <name> <url>` for a remote server, or `claude mcp add <name> -- <command>` for a local one. A connector added in the Claude app: give the human the steps from the official help page, as written there. No official page or steps: mark it ASK VENDOR and use UPLOAD for that need.
5. Install one at a time. Make one small read only test call and show the result. Then go to the next tool.
Rules: use only the human's accounts, keys and credits. Never ask for a key in chat. Start every tool read only.
<trap>A free scraper that logs in to sites, rotates proxies to get round blocks, or sends your target list to a third party. Leave it out. It breaks site terms and leaks the deal.</trap>
Stop: no official page means do not install it. Mark that need UPLOAD and carry on. Every other unit still runs.
Output: the table, what was installed, each test result, and the handoff. Next: frame-the-claims.
</unit>
Click to copy
<unit name="frame-the-claims">
Follow the shared rules in prompt 01.
What it does: turns the CIM into a short list of testable claims, ranked by how much of the price rests on each.
Use when: first, before any outside search.
Needs: the CIM or teaser, the human's thesis and worry from prompt 01.
Method:
1. List every claim about growth and its driver, retention, pricing power, market position, pipeline, hiring plan and product. Quote each word for word with its page. Label each MARKETING.
2. Keep the 4 to 6 claims the valuation rests on. A claim the model or the multiple depends on ranks first. Put the claims tied to the human's thesis and worry at the top.
3. For each claim, write what you would expect to see from outside if it is true, and what you would see if it is not.
4. Note every number the claim needs from the data room (ARR bridge, cohort file, price list, headcount by month).
<trap>Testing what is easy to find instead of what carries the price. Twenty claims checked lightly is worse than five checked properly.</trap>
Stop: no CIM and no claims typed out: build the list from the website and press releases, labelled MARKETING, and say the list may miss what the seller tells buyers.
Output: a claims table (claim · page · weight in the price · true if · false if · data room number needed), then the handoff. Next: source-map.
</unit>
Click to copy
<unit name="source-map">
Follow the shared rules in prompt 01.
What it does: matches each claim to the public signals that bear on it, with the known bias of each signal.
Use when: after frame-the-claims.
Needs: the claims table, {{TARGET}}, {{COMPETITORS}}, {{PERIOD}}.
Method:
1. For each claim, pick from: archived pricing pages, careers page and job boards, product review sites, app store reviews and ranks, employee review sites, web traffic estimates, search interest, the target's customer logos and case studies over time, public customers' filings and press, competitors' same signals.
2. For each pair write the bias: who the signal undercounts, how noisy it is, and how far back it goes.
3. Mark any claim no public signal can test as NOT TESTABLE FROM OUTSIDE, and name the data room document that tests it.
4. Search each source once to confirm it exists for this target. Record the URL and the date read. A source with no coverage is NOT REACHED.
<trap>Using one signal for a claim it cannot test. Traffic cannot prove retention. Reviews cannot prove growth.</trap>
Stop: a source behind a login or a paid panel the human has not connected: mark it NOT REACHED and carry on with the rest.
Output: the source map (claim · signal · URL · coverage dates · bias), then the handoff. Next: pricing-history.
</unit>
Click to copy
<unit name="pricing-history">
Follow the shared rules in prompt 01.
What it does: rebuilds list prices over the period from archived pages, then splits the seller's growth into price and volume.
Use when: any claim about growth, pricing power or "no price increase".
Needs: the pricing page URL, the CIM growth figure and retention figure, the CIM's renewal terms.
Method:
1. Open the Wayback Machine listing for the pricing page (web.archive.org/*/<the pricing page URL>). Read one capture per quarter across the period. Record each plan, its price, its unit and the capture date.
2. Price change per plan: new price / old price - 1. Note changes to the unit (per seat, per employee, per transaction), which change price without changing the number.
3. If renewals move to list price: prior revenue = current revenue / (1 + growth rate), on the seller's own figures. Retained base = prior revenue x gross retention. Price effect = retained base x price change. Churn = prior revenue - retained base. New business = current revenue - (retained base + price effect). Show every line.
4. If the captures do not show the month of the change, assume the whole base repriced, say the price effect may be overstated, and ask management for the renewal schedule.
5. Compare the split with the CIM's stated driver.
<trap>Reading headline growth as volume when the list price rose and renewals moved to it. A 25% price rise inside 40% growth leaves far less new logo growth than the CIM says.</trap>
Stop: no captures for the pricing page, or the page says "contact sales" for the whole period: mark the split OPEN and name the data room price list. The hiring, voice and demand units still run.
Output: a price history table, the growth split with its arithmetic, a verdict on the claim (SUPPORTED, CONTRADICTED or NOT TESTABLE FROM OUTSIDE) with its tag, then the handoff. Next: hiring-signal.
</unit>
Click to copy
<unit name="hiring-signal">
Follow the shared rules in prompt 01.
What it does: counts open roles by function and month, and reads them against the plan.
Use when: any claim about growth, expansion, a new product or a new market.
Needs: the careers page URL, archived captures of it, public job board listings for {{TARGET}}.
Method:
1. Count open postings by month and by function: sales, customer success, engineering, product, finance, operations. Count each role once, even if it is posted on several boards.
2. Read the locations and seniority. A new country or a first senior hire in a function says more than the count.
3. Set the mix against the plan. A growth plan needs sales and customer success roles. A product plan needs engineering roles.
4. Note roles that stay open for many months, or are posted, pulled and posted again.
<trap>Counting postings as headcount. Postings are intent to hire. They never prove the team grew.</trap>
Stop: no careers page and no job board coverage: mark hiring NOT REACHED. Ask management for headcount by function by month. The other units still run.
Output: a postings table by month and function, three findings, each tagged, then the handoff. Next: customer-voice.
</unit>
Click to copy
<unit name="customer-voice">
Follow the shared rules in prompt 01.
What it does: reads public product reviews, app store reviews and employee reviews for themes, churn reasons and competitor names, by period.
Use when: any claim about retention, satisfaction, product strength or team.
Needs: the review pages for {{TARGET}} and {{COMPETITORS}}, or exports of them.
Method:
1. Group reviews by quarter. Record the count and the average rating per quarter.
2. Code each review into themes: price, support, reliability, missing feature, ease of use, switching to a named competitor. Quote two short lines per theme with their dates.
3. Mark any quarter where a theme jumps, and check it against a dated event (price change, launch, outage, reorg).
4. For employee reviews, look at sales and support roles apart from the rest.
<trap>Reading a handful of angry reviews as churn. Under 30 reviews in a period says nothing on its own, and unhappy users write more.</trap>
Stop: fewer than 30 reviews in total: report the themes and say the sample is too small to tag anything. The demand and logo units still run.
Output: a themes table by quarter, the competitor names users switch to, findings with tags, then the handoff. Next: demand-signal.
</unit>
Click to copy
<unit name="demand-signal">
Follow the shared rules in prompt 01.
What it does: reads the direction of demand over the period against the seller's growth claim.
Use when: any claim about demand, growth or market share.
Needs: web traffic estimates, search interest and app ranks for {{TARGET}} and {{COMPETITORS}}, from tools the human has connected or exports they paste.
Method:
1. For each signal, compare the same month year on year: current month / same month last year - 1.
2. Mark break points: a step up or down that lasts three months or more. Match each to a dated event.
3. Compare the direction with the growth the CIM claims. Agree, disagree or cannot tell.
4. Do the same for each competitor, and show which way relative share is moving.
<trap>Reading the traffic level as revenue, or a seasonal peak as a new run rate.</trap>
Stop: no traffic, search or app data reached: mark demand NOT REACHED and say which tool would answer it. The other units still run.
Output: a year on year table per signal, the break points with their events, the direction against the claim, each tagged, then the handoff. Next: customer-base.
</unit>
Click to copy
<unit name="customer-base">
Follow the shared rules in prompt 01.
What it does: tracks the named customers on the target's site over time, and looks for public customers who name the vendor.
Use when: any claim about logos, enterprise wins, concentration or retention.
Needs: the target's customer, case study and homepage URLs and their archived captures, plus the CIM's top customer claims.
Method:
1. List every logo and case study per capture, by quarter. Mark each logo added and each logo removed, with the capture dates.
2. For each removed logo, look for a second source: the customer's own press, a competitor's case study naming them, a job posting at the customer that asks for the competitor's product.
3. For named public customers, search their filings and press for the vendor's name.
4. Compare with the CIM's claims about wins, concentration and retention.
<trap>Calling a removed logo churn. Sites drop logos in a redesign, or when a customer asks. It needs a second source.</trap>
Stop: no customer pages or captures: mark customer base NOT REACHED and name the data room customer list. The other units still run.
Output: a logo timeline (added, removed, second source found or not), findings with tags, then the handoff. Next: competitor-read.
</unit>
Click to copy
<unit name="competitor-read">
Follow the shared rules in prompt 01.
What it does: puts the target next to {{COMPETITORS}} on price, hiring, reviews and demand, to test the market position claim.
Use when: any claim about share, leadership or winning against named rivals.
Needs: the outputs of units 06 to 10, and the same public sources for each competitor.
Method:
1. One row per company: list price for the closest plan, open sales roles, review rating and count for the latest quarter, year on year demand.
2. Mark where the target leads, trails or is level, with the source for each cell.
3. Note which competitor users name when they switch (from customer-voice).
4. Set the result against the CIM's position claim.
<trap>Calling a leader from one signal. A company can lead on traffic and lose on price and reviews.</trap>
Stop: a competitor with no public data: leave its row blank with NOT REACHED. The comparison still runs on the rest.
Output: the side by side table, a verdict on the position claim with its tag, then the handoff. Next: outside-in-gate.
</unit>
Click to copy
<unit name="outside-in-gate">
Follow the shared rules in prompt 01.
What it does: an independent check. It redoes the key figures from the raw evidence BEFORE it reads the earlier tables, then blocks the scorecard if they do not tie.
Use when: after all earlier units, before the scorecard.
Needs: the raw evidence (captures, counts, ratings, estimates with their URLs and dates) and the CIM pages. Do not read the earlier outputs until step 3.
Method:
1. From the raw evidence only, recompute: each price change, the growth split, the year on year demand figures, the postings counts by function, the review counts and ratings by quarter.
2. Check every source line: URL opens, date read is recorded, label is right. A figure that traces to a numbered evidence row or page with no URL is a source note, not a BLOCK.
In one chat session, do the recompute in a fresh message before you scroll back to the earlier tables. In Claude Code, run the gate as a separate session on the raw evidence folder.
3. Now compare with the earlier tables. Any figure that differs: show both workings and the cause.
4. Check every flag against the normal list in prompt 01 again. Close any that match, with one "Checked, normal" line.
<trap>Checking the earlier tables against themselves. The gate only counts if it starts from the raw evidence.</trap>
Stop: a figure that does not tie and cannot be traced is a BLOCK on the scorecard line that uses it. Do not produce that line until the named person resolves the gap. The rest of the scorecard still runs, and the blocked line says BLOCK and why.
Output: a verdict of PASS or BLOCK, a tie out table (figure · earlier · re-derived · ties or not), closed flags, blocked lines, then the handoff. Next: scorecard-and-memo.
</unit>
Click to copy
<unit name="scorecard-and-memo">
Follow the shared rules in prompt 01.
What it does: writes the outside-in scorecard and a one page memo section for the investment committee, and a one file HTML view of the scorecard from the human's own numbers.
Use when: last, after the gate passes.
Needs: the gate output, the claims table, the human's thesis and worry from prompt 01.
Method:
1. One row per claim: claim and page · verdict (SUPPORTED, CONTRADICTED, NOT TESTABLE FROM OUTSIDE, OPEN) · the evidence in one line · tag · the question for management.
2. Only CHANGES THE DECISION rows reach the memo's view. WORTH A QUESTION rows become the management question list. EXPLAINED BY CONTEXT rows go in a short "Checked, normal" list.
3. Write the memo section: the view in two sentences, what it means for the thesis and the worry, the price relevant gaps, and what happens by default if nobody acts on the open items.
4. Write one HTML file with the scorecard table, colour by verdict, every cell linked to its source. No outside scripts or fonts.
<trap>A memo that lists everything found. The committee needs the two or three things that move the price and what to ask about them.</trap>
Stop: the gate blocked a line: the memo shows it as BLOCK with the reason, and the view says what it cannot yet conclude.
Output: the scorecard, the memo section, the management question list, the HTML file, then the handoff. Sign-off: the named person signs before it leaves the deal team.
</unit>

From a working build to an operating system.

For PE deal teams and corp dev: test every CIM claim against pricing history, hiring, reviews, demand and customer logos, each figure tied to its public source.

Path A · free

Run the build independently

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.

Back to the prompts
Path B · done with you

Connect it to your operation

We would deliver a private deal room: the review loaded into your Claude account, read only connectors to the traffic, review and VDR data your firm already licenses, a monthly refresh while the deal runs, and a sign off step so no untied figure reaches IC.

Book a build call
data safety

Before you use live numbers

  • • 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

Review authored by consultance.ai. It uses public pages only and the reader's own accounts; respect each site's terms and never contact the target's staff or customers. Your deal data stays in your own Claude account; we never see it. A named human owns the IC view.

Bring this workflow into your operation.

Book a build callBack to the library

Want this wired into your stack instead of running it yourself? That is our AI deal desk and finance automation service.

Newsletter

New builds, as they ship

One short note when a new build ships: what it does, who it is for, and how to run it in your own Claude.

Free. Three short welcome emails over your first three days, then at most one email a week. Unsubscribe in one click.

More like this

Related builds

All builds
Finance and data
Finance and data

GP Fundraise Prep Pack

For emerging manager GPs raising a fund: an LP target list ranked by room this vintage, a brief for every LP meeting, and track record pages tied to the fund's own records.

Setup guide
Finance and data
Finance and data

Insurance Agency Diligence Review

For agency acquirers and PE in insurance distribution: break down the book, split growth into lost, rate and new, tie commissions to carrier statements, model the earnout.

Setup guide
Finance and data
Finance and data

LP Due Diligence Questionnaire Desk

For GP investor relations and fund CFOs: match a new LP questionnaire to your past answers, draft the rest, flag every stale or changed answer, and stop any figure that does not tie.

Setup guide
Get the free kitBook a call

Share with your workflow owner.

Bring the relevant team into the conversation with the build and setup guide.

Forward by email
in one line

What is Outside-In Diligence Review?

Outside-In Diligence Review is a finance and data build in the consultance.ai AI Build Library. For PE deal teams and corp dev: test every CIM claim against pricing history, hiring, reviews, demand and customer logos, each figure tied to its public source. It fits PE and growth equity deal teams, corp dev, and independent sponsors who want to test the seller's story with public signals before the data room tells its side. Setup difficulty is Medium, with 4 plain-English steps.

What does Outside-In Diligence Review do?

For PE deal teams and corp dev: test every CIM claim against pricing history, hiring, reviews, demand and customer logos, each figure tied to its public source.

Who is Outside-In Diligence Review for?

It fits PE and growth equity deal teams, corp dev, and independent sponsors who want to test the seller's story with public signals before the data room tells its side.

How hard is Outside-In Diligence Review to set up?

Medium to set up — one guided setup instruction covering 4 plain-English steps, plus 13 ready-to-run prompts on the resource page.

How would consultance.ai build this out?

We would deliver a private deal room: the review loaded into your Claude account, read only connectors to the traffic, review and VDR data your firm already licenses, a monthly refresh while the deal runs, and a sign off step so no untied figure reaches IC.

What are the licensing terms?

Review authored by consultance.ai. It uses public pages only and the reader's own accounts; respect each site's terms and never contact the target's staff or customers. Your deal data stays in your own Claude account; we never see it. A named human owns the IC view.

Connect this build to your workflow.

Discuss how Outside-In Diligence Review fits your systems, controls, and team. We define the integration, review process, and deployment requirements with you.

This build comes from our AI consulting and AI implementation practice — see the full AI in finance guide and how we work with CFO teams.

Book an audit