Your team has the Finance & Ops playbooks — the worked systems for cleaning an export, reconciling two files, auditing spend, modeling cash, running the close, and briefing leadership. Those tell people what to do. This guide is the layer underneath: the operating model that keeps the work accurate, defensible, auditable, and compliant while they do it. It’s written for a finance leader rolling Claude out across a team, not for a solo analyst experimenting.
The single mistake that turns an AI rollout into an incident isn’t a clumsy prompt — it’s the absence of a rule about financial data, verification, and who signs off on a number. None of what follows is heavy process. It’s the handful of guardrails that let your team move fast because the boundaries are clear.
The operating model: Claude computes, a human owns the number
Start with the one principle everything else hangs off: Claude computes; a human owns every number. It is the fastest junior analyst you’ve ever had — tireless on cleaning exports, reconciling files, totaling with shown work, modeling, and drafting summaries — and it has no judgment, no accountability, and a habit of stating wrong numbers with total confidence.
So the division of labour is fixed:
- Claude does the grind: profile the messy export, reconcile the accounts, total with the formula shown, flag the anomalies, project the runway, draft the leadership summary, assemble the board pack.
- A person does the verifying and the deciding: spot-check the math, confirm the reconciliation ties out, decide what’s a write-off versus a chase, sign off, and own what gets reported.
A team that internalizes this gets the speed without the risk. A team that forgets it reports a confident wrong total, or closes the month on a reconciliation that was forced to balance. The playbooks are all built around this line — they take you to a verified result fast and hand the judgment back to a human on purpose. And one rule sits above all of them: Claude computes and drafts; your accounting system stays the system of record.
What financial data is safe to share — and what isn’t
On Desktop, the folder is the boundary — Claude only sees what you open — and the “Ask permissions” prompt asks before each read. That makes the data rule simple to state and easy to follow:
- Safe to work with (inside an approved workspace): your own exports, reconciliations, category lists, budgets, cash-flow models, and summaries. These are the raw material of the playbooks, and working from a local copy is fine.
- Handle with extra care / strip first: bank account and card numbers, payroll and individual compensation detail, customer PII on invoices, and anything under NDA. Use a
[placeholder]for anything truly private in a prompt you share, and strip an export to what the task actually needs before it reaches the chat.
Two habits make this automatic. First, drop in only what the task needs — a spend audit runs on vendor, amount, and date, not on full account numbers. Second, when an export does carry sensitive detail, keep the whole task inside the workspace your company has approved for that data, rather than a general chat. The messy-export, reconciliation, and board-pack playbooks all flag this at the exact step it matters.
The verification and sign-off gate
The fastest way to lose trust in finance isn’t a slow report — it’s a confident wrong one. So put one gate between a computed number and a reported number, and make it explicit. This is the finance equivalent of marketing’s brand-and-legal gate or sales’ deal-desk:
- Show the work, then check it. Every total, reconciliation, and projection comes with its formula or steps, and a human spot-checks one figure by hand against the raw file. If it ties, the method is trustworthy across the rest. A number with no shown work doesn’t get reported.
- Reconciliations tie out; gaps get flagged, never forced. A reconciliation that doesn’t balance is a finding, not a failure — the discrepancy is exactly what you needed to surface. Never let Claude paper over an unexplained difference to make the numbers look clean.
- Traceability for anything that leaves the team. A leadership summary or board pack carries an appendix where every headline figure traces to a source file and a formula. When someone asks “where’s that number from?”, the answer is one scroll away — that’s what makes it defensible to a board or an auditor.
This isn’t bureaucracy bolted on — it’s a single named step in the workflow. The reconciliation playbook insists the report ties out; the leadership-summary and board-pack playbooks put the verified appendix under the narrative, because at enterprise scale what gets a finance team in trouble is an unverifiable number, not a slow one.
Who owns what (a light RACI)
Speed dies in ambiguity about who can sign off on a number. You don’t need a formal RACI matrix — you need four roles named for any piece of finance work, and they should sit with different people where the control matters:
- Prepares — the analyst running the playbook with Claude.
- Reviews — a second person who checks the reconciliation ties out and spot-checks the totals.
- Approves — whoever can sign off: a controller on the close, the CFO on what goes to the board, the budget owner on a variance.
- Owns — a person owns the reported number, can answer for it, and decides the judgment calls (write-off vs chase, capex vs opex).
The point is that Claude is never any of these four. It’s the tool the Prepares role uses. Crucially, keep the preparer and the approver separate for anything that moves money or hits the books — the person who reconciles an account shouldn’t be the one who unilaterally writes off its exceptions. Write these four names down once per workstream (close, budget, collections) and most “who signed off on this?” fire drills — and most control gaps — disappear.
Reporting in two languages: Arabic and English
For a MENA team, bilingual isn’t a translation step bolted on at the end — it’s a parallel track. The rule that matters: adapt the words, share the numbers.
- Author each narrative in its own language. A leadership summary or board narrative for an Arabic-reading audience is written in Arabic from the start, for the right register and formality — not translated from the English. A machine-translated financial briefing reads translated, and in this region that undercuts the credibility the numbers earned.
- One shared numbers appendix. The prose differs by language; the figures must not. Keep a single verified appendix both versions reference, so the Arabic and English packs can never quietly disagree on a total.
- Same gate, both languages. The Arabic summary clears the same verification and sign-off — and a fluent reviewer checks the language. Don’t let the second language skip the review because nobody on the approval chain reads it; find someone who does.
Treated this way, Arabic is a first-class peer of English in your reporting, not an afterthought — which, in this region, is the difference between a board that trusts the pack and one that quietly doesn’t.
The capability path: from a clean export to a board-ready pack
The playbooks are arranged as a staged path, not a flat menu, because the later ones genuinely depend on the earlier ones. Run a team through it in order:
- Stage 1 — Foundations. Get your data and your rules trustworthy: the habit of making a messy export trustworthy before you trust a total, and a categorization rulebook (
categorization-rules.md) so the same vendor lands in the same bucket every month. Skip these and every total, audit, and close starts from dirty data and a guess. - Stage 2 — Repeatable systems. Turn the foundations into monthly engines: reconcile two files, audit the month’s spend, track budget vs. actuals, consolidate vendor & SaaS spend, run AR aging and collections, build a cash-flow model, and brief leadership.
- Stage 3 — Orchestration. Run the big recurring moments that pull everything together: the full month-end close and the board & investor pack — where the close makes the numbers right and the pack makes them presentable and defensible.
On the Finance & Ops team hub the path is laid out stage by stage with a progress bar, and your team can mark each playbook done as they go — so “we’re upskilling the team” becomes a number you can actually see, not a hope. (Progress is tracked on each person’s device; a shared, manager-level team view is the natural next step once you’ve run the path.)
Rolling it out: a 30/60/90 plan
Don’t roll everything out to everyone on day one — that’s how rollouts fizzle. Stage it the same way the path is staged. Here’s the whole arc as a checklist you can copy into your own doc:
30/60/90 — rolling Claude out across a finance & ops team
First 30 days — Foundations
- Pick 2–3 willing people across finance and ops.
- Build the two foundations: the trustworthy-export habit and a
categorization-rules.md the whole team shares.
- Agree the one-page guardrail: what financial data is safe to share, and
the verification rule — show the work, spot-check, tie out, sign off.
- Each person runs their first real Stage 1 playbook end to end.
Days 30–60 — Repeatable systems
- Put the monthly engines on Claude: reconcile, spend audit, budget vs
actuals, vendor review, AR collections, the cash-flow model, the summary.
- Capture the prompts that work into a shared CLAUDE.md and slash commands.
- Name the four roles per workstream: prepares / reviews / approves / owns —
keeping preparer and approver separate where it moves money.
Days 60–90 — Orchestration
- Run one real month-end close and one board pack end to end through the
playbooks.
- Stand up the controls (segregation of duties, the verified appendix) as
named runbook steps, not afterthoughts.
- Review the capability path: who's completed which stage, where the gaps
are, and which seats to add next.The urge to mandate it for everyone at once is the urge to resist. Adoption is a habit that spreads, not a switch you flip. Start a few people on the foundations, give them the two source habits and a one-page rule, let them capture what works — and by day 90 you have a team that’s run a real close and a board pack through the system, a path you can point to, and the proof you need to widen it. The general team rollout playbook goes deeper on the cross-role mechanics when you’re ready to scale it past finance.