For CFOs defending AI spend to the board: price every AI finance agent against the salary it replaces and prove its output. 13 prompts and a live calculator replace your FinOps spreadsheet with a one-page Agent P&L.
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.
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.
Free account, no card, 30 seconds. This tab stays open so you can come back.
Open claude.ai ↗Claude reads this page, asks which computer you are on, then guides you step by step until it works. If anything errors, tell Claude what you see, and it fixes it with you.
I want to build my Agent P&L with the Agent P&L Board Pack on my own numbers. 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), plus one calculator that opens in a browser tab. Treat me like a CFO or finance lead 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 seat spend, GL, and forecast, and produce a board-facing Agent P&L that gives every AI finance agent a cost line AND an output line. Nothing installs on my machine. Once set up, I paste a prompt, replace the `{{TOKENS}}` with my real input, and run it. A finance owner reviews before anything goes to the board.
Walk me through this one step at a time. Wait for me to confirm each step before moving on:
1. First, no login needed: open the live Agent P&L calculator from the button on this page in a new browser tab. Price the six default agents against the salary each replaces, set the honest output share per agent, enter your real monthly seat spend, and read your net Agent P&L. This is the shape of what the prompts will make defensible on your own data. Come back when you have a number.
2. Open a browser and go to `claude.ai`, or open the Claude desktop app. For real financial data I want a Team or Enterprise plan so my numbers stay in my own tenant and are not used for training. Tell me how to confirm that setting. If I prefer to work in Excel, tell me how to enable the Claude add-in from the Excel ribbon instead.
3. In the left sidebar, click "Projects", then "Create Project". Name it "Agent P&L" or the entity name.
4. In Project Knowledge, add the data the prompts work from (the source I pick in prompt 01): my AI seat and usage export, GL or trial balance, AP ledger, and my current forecast, as CSV or PDF. If I use the Excel add-in instead, tell me to keep the workbook open and skip the upload.
5. In the Project's Custom Instructions box, paste this: "You are a finance leadership team supervised by a CFO. Every figure ties to a source or my input. Never invent a number. No AI agent goes on the P&L with a cost line and no evidenced output line. Flag everything for human sign-off before it leaves finance."
6. Open a new chat inside the Project. Copy prompt 01 (the Agent P&L onboarding router) from the vault on this page. Paste it. Answer its five blocks (stage, where your data lives, which agents, context, output bar). This configures every later prompt.
7. Run prompts 03, 04, 05, then 06 in order. This builds the salary-replacement basis, the output attribution ledger, the agent cost model, then the Agent P&L itself. Replace each `{{TOKEN}}` with my real input. Now the calculator's numbers from step 1 are backed by my own data.
8. Run prompt 07 (reconcile the Agent P&L). It re-derives the whole P&L a second, independent way from the raw inputs and STOPS if the two do not tie. Tell me plainly if it returns a MISMATCH, and do not move on until it returns RECONCILED. A board number that only checks against itself is not checked.
9. Run prompt 13 (board one-pager). Read it. Confirm every agent row shows BOTH a cost line and an output line, every figure ties to a source and to the reconciled total, and nothing is presented as a layoff when it is really output attribution.
10. Before the pack goes to the board, run bonus prompt B2 (the science-project killer). Any agent with a cost line and no evidenced output line gets marked "watch" or cut, not carried.
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, `<review_gate>`, `{{TOKEN}}`, output attribution, cost avoided.
- If the UI looks different (web vs desktop vs 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 straight into a chat.
- Never tell me to install anything via Terminal. The whole bundle runs from a browser tab, the desktop app, or Excel.
- Anti-pattern call-outs: if I say "cost avoided is not real cash", agree and coach me to present it as capacity freed, not cash saved. If I say "my IT blocked claude.ai", tell me about the Team/Enterprise plan with SSO or the desktop app. If I say "I do not have a paid plan", tell me the free plan tests the first prompt but real financial work needs the firm tenant.
First message: ask me only "Are you working in claude.ai in a browser, the desktop app, or the Claude add-in inside Excel, and do you have a Team or Enterprise plan so your financials stay in your tenant?" Then start with step 1.
When prompts 01 through 06 run cleanly, prompt 07 returns RECONCILED, and prompt 13 produces a board one-pager you would actually present, switch into "DM mode" and ask me which agent's output line is weakest so I can help wire that one to live data.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.
Browser-based. No signup. Drop in your numbers and see the trade in real time. Opens in a new tab so the prompts stay where you left them.
For CFOs defending AI spend to the board: price every AI finance agent against the salary it replaces and prove its output
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 ↑We would stand up the governed agent stack inside your own Claude tenant: each finance agent wired to your data, an output attribution ledger logged every run, model and spend guardrails set, human sign-off gates enforced, and the board Agent P&L regenerated on a monthly schedule.
Book a build call →Cost avoided is output attribution, not a layoff or guaranteed cash. Public salary benchmarks are illustrative, tune them to your market. A qualified finance owner reviews every output before it reaches the board. This is not accounting, audit, or legal advice.
Want this wired into your stack instead of running it yourself? That is our AI deal desk and finance automation service.
The AI tools, launches, and shifts that actually matter, in plain English. New library drops the moment they land.
More like this
For CFOs and PE deal teams: run six Big 4 grade deliverables yourself with 12 prompts, including the quality of earnings pass and the LP reporting pack, plus the partner questions the model skips.
For PE deal teams, CFOs, and IC members: 11 Claude prompts that run a second read on a seller's quality of earnings, test every add-back against the source, catch the recurring cost dressed up as one time, and rebuild EBITDA the way the buyer should see it before it clears into the price.
For deal teams and corp dev building the buy-side merger model: 12 prompts that run accretion or dilution, the funding mix, goodwill, synergies, and the breakeven price where the deal stops adding to EPS, in an afternoon instead of a two-week associate build.
The person drowning in this every week is the one who'll actually want it.
AI Agent P&L Board Template is a finance and data build in the consultance.ai AI Build Library. For CFOs defending AI spend to the board: price every AI finance agent against the salary it replaces and prove its output. 13 prompts and a live calculator replace your FinOps spreadsheet with a one-page Agent P&L. It fits CFOs, controllers, and finance leaders at $5M-$500M companies and PE portfolio companies who bought AI seats and cannot yet tie them to a dollar of output. Setup difficulty is Medium, with 5 plain-English steps.
For CFOs defending AI spend to the board: price every AI finance agent against the salary it replaces and prove its output. 13 prompts and a live calculator replace your FinOps spreadsheet with a one-page Agent P&L.
It fits CFOs, controllers, and finance leaders at $5M-$500M companies and PE portfolio companies who bought AI seats and cannot yet tie them to a dollar of output.
Medium to set up — one guided setup instruction covering 5 plain-English steps, plus 13 ready-to-run prompts on the resource page.
We would stand up the governed agent stack inside your own Claude tenant: each finance agent wired to your data, an output attribution ledger logged every run, model and spend guardrails set, human sign-off gates enforced, and the board Agent P&L regenerated on a monthly schedule.
Cost avoided is output attribution, not a layoff or guaranteed cash. Public salary benchmarks are illustrative, tune them to your market. A qualified finance owner reviews every output before it reaches the board. This is not accounting, audit, or legal advice.
AI Agent P&L Board Template is the starting point. On a free AI audit we map where it fits your stack and what consultance.ai would build around it.
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.