HomeResourcesFinanceAndrew PershJune 30, 20269 min read

Six-Stage Finance Operating System: Claude Skills for Analysts

The six-stage finance operating system structures every deal and capital decision the same way a disciplined finance desk does: Frame the question, Analyse the sector, Value the range, Execute the plan, Control the position, and Communicate the answer. These are the Claude prompts that run each stage.

30,000+ consultants, bankers, private equity professionals

Free skills and prompts for Claude and strategy work

Templates for Claude, ChatGPT and Perplexity — from diagnostics to board-ready decks.

Six-stage finance operating system: Frame, Analyse, Value, Execute, Control, Communicate

What the six-stage finance operating system is

Most finance teams use Claude as a calculator or a summariser. They paste in a spreadsheet and ask for a number. The six-stage finance operating system uses Claude differently: it runs Claude through the same six stages a top deal team works, every time, producing consistent, defensible output from the first frame to the final memo.

The backbone is six stages: Frame, Analyse, Value, Execute, Control, and Communicate. Each stage produces a specific deliverable. The deliverable from one stage becomes the input to the next. The system prevents the most common analytical failure in finance: building a precise model against the wrong question.

The stages are not strictly linear in practice. Valuation and analysis often run in parallel. The Frame stage may repeat after new information arrives. But the logical dependencies are fixed: you cannot build a defensible valuation range without a sector map, and you cannot write a credible IC memo without a stress-tested position. For the complete finance skills system covering FP&A, treasury, and reporting, the CFO guide to Claude covers 24 domain skills built on the same architecture.

Stage 1 of 6

Frame: Lock the decision before the analysis begins

Most finance work fails before the first spreadsheet opens. The team answers the wrong question with great precision. The Frame stage prevents this. Claude locks the work to a single decision: invest, hold, exit, or restructure. It names the three assumptions the answer rests on and lists the five facts needed to test them. Everything that follows works against this brief.

Use when

Use at the start of any deal, review, or capital allocation decision, before any data is pulled.

Output

A single decision statement, three load-bearing assumptions, five facts to test, and a list of noise to ignore.

Claude prompt

Frame the decision for [company/situation]. We need to decide: invest, hold, exit, or restructure. Define the decision in one sentence. Name the three assumptions the answer stands or falls on. List the five pieces of evidence we need to test each assumption. Flag the noise that is already in the room but will not change the conclusion.

Stage 2 of 6

Analyse: Map where value sits and how the cycle could break the thesis

The Analyse stage builds the sector context the valuation must rest on. Claude maps the sector structure, runs comparables, and screens counterparty and credit quality. The output is not a summary of what is known. It is a structured read of where value sits in the sector and which forces could break the thesis before the model gets to prove it.

Use when

Use when entering a new sector, initiating coverage, or stress-testing an existing thesis against sector shifts.

Output

A sector brief covering size, demand drivers, and the cycle, a comparable screen by margin and multiple, and a counterparty and credit risk summary.

Claude prompt

Analyse [sector/company] across three dimensions. First, map the sector: size, sub-segments, the five key players and their approximate positions, primary demand drivers, and the one structural force most likely to shift the cycle in the next 18 months. Second, run a comparable screen: identify eight to twelve true peers and show where [company] sits on revenue growth, EBITDA margins, and EV multiples. Flag outliers and document why they are excluded. Third, assess counterparty and credit quality: name the top three concentration risks, the credit metric most under pressure, and the one early-warning indicator to watch monthly.

Stage 3 of 6

Value: Build the range, not a single point estimate

The Value stage produces three valuation outputs that triangulate to a defensible range: a DCF with explicit sensitivities, an LBO with returns at multiple exit scenarios, and a scenario build that shows the probability-weighted equity value. Claude documents which methodology should carry the most weight in this specific situation and why. The output is a range with a view, not a number without a story.

Use when

Use before any IC presentation, board paper, or public statement that requires a defensible valuation range.

Output

A DCF with WACC and terminal growth sensitivity table, an LBO with IRR and MOIC at three exit multiples, and a scenario-weighted equity value range.

Claude prompt

Value [company] across three methodologies. Build a DCF with a five-year explicit forecast, a terminal value using stable-growth assumptions, and a sensitivity table varying WACC by 100 basis points and terminal growth by 50 basis points. Layer in an LBO at [x]% entry leverage and show IRR and MOIC at exit multiples of [a], [b], and [c]. Run three scenarios: base, upside, and downside, with a one-paragraph narrative and probability weight for each. Show the probability-weighted equity value and document which methodology should anchor the range and why.

Stage 4 of 6

Execute: Turn the thesis into something a committee can actually run

A thesis without an execution plan is an opinion. The Execute stage turns the investment view into a deal structure with covenants, a 90-day plan with named owners, and a dependency map that shows which single assumption most threatens the plan in the first quarter. Claude writes the plan a committee can vote on, not a slide with intentions.

Use when

Use when moving from investment decision to implementation, or when structuring the terms and first-100-day plan for a new position.

Output

A deal structure with equity and debt split, covenant schedule, and governance terms, plus a 90-day plan with three to five initiatives and named KPIs.

Claude prompt

Turn the investment thesis into an execution plan for [company]. Design the deal structure: equity and debt split, three to five key financial covenants with current headroom, and the governance terms the investment requires. Draft a 90-day value creation plan with three to five initiatives. For each initiative, name the accountable owner, the target outcome, and the KPI that confirms it is on track by day 90. End with the single execution dependency that most threatens the plan in the first quarter and what would trigger a re-plan.

Stage 5 of 6

Control: Pressure-test the position before the market does

The Control stage stress-tests the position with numbers, not narratives. Claude runs a downside stress test, builds a KPI scorecard with red and amber thresholds, and sets exposure limits on the three largest concentration risks. The output is a monitoring framework that tells the team exactly when to act, not a qualitative list of things that could go wrong.

Use when

Use after a position is taken or a plan is approved, to build the monitoring framework before the first reporting period.

Output

A stress test showing covenant headroom by quarter in the downside case, a five-KPI scorecard with green, amber, and red thresholds, and exposure limits on the three largest risks.

Claude prompt

Pressure-test the position in [company]. Run a stress test with EBITDA at 20% below base: calculate covenant headroom quarterly over the forecast period and flag the first quarter where headroom falls below 15% of the threshold. Build a five-KPI monitoring scorecard with a specific green, amber, and red threshold for each KPI. Set a maximum exposure limit for the three largest concentration risks and the action triggered if the limit is breached. End with the two leading indicators that would signal the thesis is breaking six months before it appears in the reported numbers.

Stage 6 of 6

Communicate: Build the answer-first story a sceptical room cannot poke holes in

The Communicate stage produces the investment committee memo, the hostile Q&A preparation, and the board-ready recommendation. Claude writes answer-first throughout: the recommendation on the first line, the supporting argument in the body, the risks addressed before they are raised. The output earns a decision from a sceptical room rather than prompting another round of questions.

Use when

Use when preparing for IC, board, or credit committee, or when writing the formal investment memo or board paper.

Output

A three-page investment committee memo with recommendation on the first line, and a hostile Q&A covering the five hardest questions a sceptical committee would ask.

Claude prompt

Write the investment committee memo for [company/deal]. Lead with the recommendation on the first line: we recommend [invest / hold / exit / restructure] at [terms]. Structure the memo: situation and context in two paragraphs, investment thesis with three numbered supporting arguments, valuation range with the anchoring methodology explained, the three most material risks and their specific mitigants, and the ask. Follow the memo with a hostile Q&A: list the five hardest questions a sceptical committee member would ask and write the full answer to each. Keep the memo to three pages and the Q&A to one. Formal register, answer-first throughout.

Setting up the six-stage finance operating system in Claude

The six stages run inside a single Claude Project. Create one project per deal or company. Add the stage prompts as project knowledge documents. Claude carries forward context across conversations inside the same project, so you do not re-paste data between stages.

Step 1

Create a Claude Project for the deal

Go to claude.ai, open the left sidebar, click Projects, then Create Project. Name it by company or deal. This project becomes the analytical environment for the full engagement, with context persisting across every conversation you open inside it.

Step 2

Add the six stage prompts as Project Knowledge

Inside your project, open Project Knowledge and click Add Content. Paste each stage prompt as a separate document. Name each document after the stage: Frame, Analyse, Value, Execute, Control, Communicate. Claude reads them automatically in every conversation inside that project.

Step 3

Run stages in sequence, referencing by name

Reference the stage name in your prompt. The Frame output becomes the brief for the Analyse stage. The Analyse output informs the Value assumptions. You can run multiple stages in a single conversation or open a new conversation per stage.

Example prompt chain

"Run the Frame stage on [company]. We are deciding whether to invest at [terms]."
"Now run the Analyse stage using the decision brief we just built."
"Run the Value stage. Use the sector context from the Analyse stage to anchor the DCF assumptions."
"Run the Communicate stage. Lead the memo with the valuation range we established."

Tip

Always run the Frame stage first, even if it feels redundant. The decision brief it produces changes how Claude interprets every subsequent prompt. It is the difference between Claude building what you asked for and Claude building what you need.

Step 4

Turn the Communicate output into board-ready slides with Oria

The six-stage OS handles the analysis. Oria handles the presentation layer. Paste the investment memo, valuation summary, or board recommendation from the Communicate stage into Oria inside PowerPoint. Oria converts it into a fully editable slide in your corporate template in two to three minutes per slide. No manual formatting. No re-typing numbers. The analysis that took the six stages to build is board-ready inside the hour.

The quality bar for six-stage finance operating system output

Each stage is designed to hold output to the standard a senior finance professional would apply before it leaves the room. The 12 checks below are the minimum bar for output from the complete six-stage sequence. For a full set of institutional-grade skills covering these same phases, see the 18 institutional-grade Claude skills for finance.

Decision locked before the first model opens

Sector context built before the valuation

Range shown, not a single point estimate

Sensitivities explicit, not assumed away

Execution plan with named owners and KPIs

90-day dependency mapped before the vote

Downside stress run before the position is taken

Covenant headroom monitored quarterly

IC memo leads with the recommendation

Hostile Q&A prepared before the room asks it

Answer-first throughout, no buried conclusions

Sources cited, assumptions stated explicitly

Frequently asked questions

What is the six-stage finance operating system?

It is a repeatable analytical sequence for finance deals and decisions: Frame, Analyse, Value, Execute, Control, and Communicate. Each stage locks a specific output before the next begins. The system prevents the most common analytical failure in finance, which is running a precise model against the wrong question.

Can I run all six stages in a single Claude Project?

Yes. Create one Claude Project for a deal or company. Add the six-stage framework as project knowledge, and reference each stage by name in your prompts. Claude carries forward context across conversations inside the same project, so the Analyse output informs the Value stage without re-pasting data.

What finance decisions does this system cover?

The system is designed for any capital decision: equity investments, credit decisions, strategic acquisitions, portfolio company reviews, and capital allocation calls. The Frame stage adapts to the specific decision type. The five stages that follow use the same backbone regardless of the asset class.

Why start with Frame before the model?

The most expensive mistake in finance analysis is building a precise model against the wrong question. The Frame stage costs almost no time and produces a decision brief that every subsequent stage works against. Skipping it is the source of most analytical rework.

How do I turn Claude outputs into board-ready slides?

Paste the investment memo, valuation summary, or board paper from Claude into Oria inside PowerPoint. Oria converts the structured text into fully editable slides in your firm's corporate template in two to three minutes per slide. The six-stage OS handles the analysis. Oria handles the presentation layer.