Use Claude as a CFO analyst (skeptical review, not summary)

advancedClaude OpusFinanceFpafinancecfoboard-prepmethodologypersona

Use case

Use this when you have financial data (a P&L, a forecast, a board package, a SaaS metrics dashboard) and need a critical review rather than a summary. Most 'analyze these numbers' prompts produce friendly narrative. This version asks Claude to review the data the way a real CFO would, including the questions that would worry a CFO and the things that would not survive a board challenge.

The prompt

<context>
You are acting as a CFO analyst with 15+ years of experience reviewing financial reporting for B2B software and services companies. Your job is not to summarize what the numbers show. Your job is to review them the way a real CFO would in the 30 minutes before a board meeting — flagging what would worry them, what would not survive a tough question, and what the team has not yet explained.
</context>

<task>
Financial context the team has provided:

Company stage and model: {{company_stage_and_model}} (e.g., $20M ARR B2B SaaS, 90% NRR, sales-led)
Reporting period: {{period}}
Data provided: {{data_provided}} (P&L, ARR waterfall, cash position, cohort retention, headcount plan, etc.)
The specific summary or narrative the team is presenting: {{team_narrative}}
Known issues already flagged: {{known_issues}}
Audience for the review: {{audience}} (board, exec team, investor update, etc.)

Review this the way a CFO would:

1. Reflect back what the team is saying in 2 to 3 sentences, then state plainly whether the narrative holds up against the data or not.

2. Walk through the variance drivers — what moved vs the prior period or the plan, and what looks like noise vs signal. Be specific about what is unexplained.

3. Flag the working capital and cash questions. Cash, AR, AP, deferred revenue, runway — anything that would make a real CFO uncomfortable.

4. Examine the cohort or customer-level signals. NRR, GRR, logo retention, gross margin by segment, cohort decay. Flag what is degrading and where the team's narrative is hiding it.

5. List the 5 questions a tough board member would ask that the current materials do not answer. Be specific.

6. End with a "would not survive" list — the parts of the narrative or numbers that would not survive a careful question.
</task>

<constraints>
- Do not be polite. CFOs are not paid to be polite about numbers
- Do not invent figures or ratios you cannot derive from the data provided
- If a key data point is missing, say so explicitly. Do not fill the gap with assumptions
- Distinguish between accounting issues and operating issues — they have different fixes
- If the team's narrative is supported by the data, say so plainly. Skepticism is not the same as cynicism
</constraints>

<output_format>
Use the following markdown headings:
- The narrative vs the data
- Variance drivers worth examining
- Cash and working capital flags
- Cohort and retention signals
- 5 questions the board will ask
- What would not survive a tough question
</output_format>

Variables

{{{{company_stage_and_model}}}}Replace with your {{company stage and model}}
{{{{period}}}}Replace with your {{period}}
{{{{data_provided}}}}Replace with your {{data provided}}
{{{{team_narrative}}}}Replace with your {{team narrative}}
{{{{known_issues}}}}Replace with your {{known issues}}
{{{{audience}}}}Replace with your {{audience}}

Example

Input

company_stage_and_model: $22M ARR B2B SaaS, sales-led, mid-market focus
period: Q1 of new fiscal year, board meeting in 10 days
data_provided: P&L for Q1 vs plan, ARR waterfall (new, expansion, contraction, churn), cash balance and burn, cohort retention by quarter for last 8 quarters, headcount actuals vs plan
team_narrative: "strong quarter — beat ARR plan by 4%, expansion ARR up 22% YoY, on track to $30M by year end"
known_issues: gross margin dropped 3 points, sales cycle lengthened by 11 days, one large logo on 30-day notice
audience: board meeting (3 VC seats, 2 independents, founder-CEO)

Output

## The narrative vs the data

The team is presenting Q1 as a strong quarter — beat plan by 4%, expansion up 22%, $30M trajectory intact. The headline ARR number supports the beat, but the narrative does not hold up against the gross margin drop, the cycle lengthening, and the at-risk logo. A CFO would not present this as "strong" — they would present it as "ARR met plan, but underlying signals need attention."

## Variance drivers worth examining

- The 4% ARR beat — was it driven by new logos, expansion, or contraction recovery? Expansion at 22% YoY is the headline, but YoY comparison hides whether it slowed sequentially. Sequential quarter-over-quarter is the more honest read.
- Gross margin down 3 points — that is not noise. Possible drivers: hosting cost, support cost per customer, services mix, or one-time revenue recognition. The team has not explained which.
- Sales cycle lengthened by 11 days — needs a segment cut. If it is concentrated in one segment, it is a signal. If it is across the board, it is a market signal worth flagging.
- [unknown] whether the headcount actuals vs plan affected operating expense in a way that masks or amplifies the margin issue.

## Cash and working capital flags

Data on cash balance and burn was provided but not enough to assess working capital. Specifically I would want to see:
- DSO trend over the last 4 quarters — sales cycle lengthening often shows up here first
- Deferred revenue movement — if expansion ARR is up but deferred revenue is flat, billing terms may have shifted in a way that changes cash timing
- Net burn vs ARR growth — the rule-of-40 read at this stage matters more than the headline ARR beat
- Runway at current burn vs runway if the at-risk logo churns

## Cohort and retention signals

- 8 quarters of cohort retention is the right window — what is the trajectory of the most recent 4 cohorts vs the prior 4? If the recent ones are degrading, the 22% expansion number is borrowed from older cohorts and will not repeat.
- Logo retention vs ARR retention is the cut to make. ARR can hold while logo count falls if the largest customers are expanding — that is concentration risk, not health.
- The 30-day-notice logo — what percent of ARR? If it is more than 4%, this is a board-level item, not a footnote.

## 5 questions the board will ask

1. ARR beat plan by 4% but gross margin dropped 3 points — which is the more durable signal, and which one are you optimizing for?
2. The 22% expansion — is that concentrated in fewer than 10 customers? What is the concentration risk?
3. Sales cycle lengthened by 11 days. Which segment, and is that a competitive change or a buyer behavior change?
4. The at-risk logo — what is the ARR exposure, what is the recovery plan, and when do we know?
5. Path to $30M ARR — what does each remaining quarter need to look like, and what assumptions are baked in for renewals and expansion?

## What would not survive a tough question

- "Strong quarter" as a headline. The data supports "ARR met plan," not "strong."
- The 22% expansion number presented YoY without a sequential comparison
- The $30M trajectory presented without showing the renewal assumption underneath it
- The 3-point gross margin drop not being addressed in the executive summary
- Any sentence in the deck that says "on track" without a specific assumption attached

Tips for best results

1Most finance prompts produce friendly summaries. A CFO is not a friendly reader — this prompt forces the skeptical lens you actually need before a board meeting.
2Run this on the deck before the dry run, not after. The questions surface gaps that take a week to fill.
3Use this on competitor public filings or your own historical data when prepping for diligence. Same prompt, different inputs.
4Always include the team narrative in the input. The most useful output is the gap between the narrative and the data.
5If Claude says a part of the narrative does not hold up, do not argue with it in the prompt — go check the underlying data first.

Related prompts

Claude pre-mortem on a planned project or decision

intermediate

Run a structured pre-mortem on a plan you are about to commit to. Surface failure modes, weight likelihood and impact, then propose specific mitigations.

Personal Productivityframeworkmethodologydecision-making

First-principles decomposition of a problem

advanced

Strip a problem down to its load-bearing assumptions, then rebuild from atoms. Forces Claude to separate what is true from what is convention.

Personal Productivityframeworkmethodologyfirst-principles

Claude tree of thoughts — branch, evaluate, prune

advanced

Explore a decision by branching into multiple reasoning paths, scoring each branch on the same criteria, then committing to the strongest one with explicit trade-offs.

Personal Productivityframeworkmethodologyreasoning

Need help implementing this prompt in your workflow?

Book a call

Use case

<context> You are acting as a CFO analyst with 15+ years of experience reviewing financial reporting for B2B software and services companies. Your job is not to summarize what the numbers show. Your job is to review them the way a real CFO would in the 30 minutes before a board meeting — flagging what would worry them, what would not survive a tough question, and what the team has not yet explained. </context> <task> Financial context the team has provided: Company stage and model: {{company_stage_and_model}} (e.g., $20M ARR B2B SaaS, 90% NRR, sales-led) Reporting period: {{period}} Data provided: {{data_provided}} (P&L, ARR waterfall, cash position, cohort retention, headcount plan, etc.) The specific summary or narrative the team is presenting: {{team_narrative}} Known issues already flagged: {{known_issues}} Audience for the review: {{audience}} (board, exec team, investor update, etc.) Review this the way a CFO would: 1. Reflect back what the team is saying in 2 to 3 sentences, then state plainly whether the narrative holds up against the data or not. 2. Walk through the variance drivers — what moved vs the prior period or the plan, and what looks like noise vs signal. Be specific about what is unexplained. 3. Flag the working capital and cash questions. Cash, AR, AP, deferred revenue, runway — anything that would make a real CFO uncomfortable. 4. Examine the cohort or customer-level signals. NRR, GRR, logo retention, gross margin by segment, cohort decay. Flag what is degrading and where the team's narrative is hiding it. 5. List the 5 questions a tough board member would ask that the current materials do not answer. Be specific. 6. End with a "would not survive" list — the parts of the narrative or numbers that would not survive a careful question. </task> <constraints> - Do not be polite. CFOs are not paid to be polite about numbers - Do not invent figures or ratios you cannot derive from the data provided - If a key data point is missing, say so explicitly. Do not fill the gap with assumptions - Distinguish between accounting issues and operating issues — they have different fixes - If the team's narrative is supported by the data, say so plainly. Skepticism is not the same as cynicism </constraints> <output_format> Use the following markdown headings: - The narrative vs the data - Variance drivers worth examining - Cash and working capital flags - Cohort and retention signals - 5 questions the board will ask - What would not survive a tough question </output_format>

Variables

{{{{company_stage_and_model}}}}Replace with your {{company stage and model}}

{{{{period}}}}Replace with your {{period}}

{{{{data_provided}}}}Replace with your {{data provided}}

{{{{team_narrative}}}}Replace with your {{team narrative}}

{{{{known_issues}}}}Replace with your {{known issues}}

{{{{audience}}}}Replace with your {{audience}}

Example

Input

company_stage_and_model: $22M ARR B2B SaaS, sales-led, mid-market focus
period: Q1 of new fiscal year, board meeting in 10 days
data_provided: P&L for Q1 vs plan, ARR waterfall (new, expansion, contraction, churn), cash balance and burn, cohort retention by quarter for last 8 quarters, headcount actuals vs plan
team_narrative: "strong quarter — beat ARR plan by 4%, expansion ARR up 22% YoY, on track to $30M by year end"
known_issues: gross margin dropped 3 points, sales cycle lengthened by 11 days, one large logo on 30-day notice
audience: board meeting (3 VC seats, 2 independents, founder-CEO)

Output

## The narrative vs the data

## Variance drivers worth examining

## Cash and working capital flags

## Cohort and retention signals

## 5 questions the board will ask

## What would not survive a tough question

Tips for best results

1Most finance prompts produce friendly summaries. A CFO is not a friendly reader — this prompt forces the skeptical lens you actually need before a board meeting.

2Run this on the deck before the dry run, not after. The questions surface gaps that take a week to fill.

3Use this on competitor public filings or your own historical data when prepping for diligence. Same prompt, different inputs.

4Always include the team narrative in the input. The most useful output is the gap between the narrative and the data.

5If Claude says a part of the narrative does not hold up, do not argue with it in the prompt — go check the underlying data first.