Purpose
Turn raw operational metrics for a single finance process into a defensible, board-ready ROI model — quantifying labor savings, error reduction, cycle time compression, and working capital gains against realistic implementation and ongoing costs, then stress-testing the result so the finance leader knows exactly what they're committing to and when it pays back.
When to Use
Invoke this skill when the user asks to:
- Build a business case for automating a finance process (AP, reconciliations, close, expenses)
- Calculate the ROI, payback period, or NPV of an RPA, AI, or software automation initiative
- Justify a finance automation software or tooling investment to the CFO or board
- Compare the cost of the current manual process against an automated future state
- Quantify the hidden cost of errors, rework, or slow cycle times in a finance workflow
- Model the working capital or early-payment-discount upside of faster processing
Inputs Required
Before starting, collect:
- Process being automated: The specific workflow (e.g., "AP invoice processing," "bank reconciliations," "intercompany close")
- Transaction volume: Number of items processed per month or year (invoices, reconciliations, journal entries)
- Fully-loaded FTE cost: Annual cost per employee including benefits/overhead, and number of FTEs (or FTE fractions) on the process
- Current handling time: Average minutes per transaction, or total hours the team spends monthly
- Current error/exception rate: % of transactions requiring rework, plus average cost or time to resolve one error
- Current cycle time: Days from receipt to completion (e.g., invoice-to-payment, close duration)
- Implementation & run cost estimate: Expected one-time build/license cost and annual ongoing/subscription cost (if unknown, note it for estimation)
If working capital inputs (payment terms, discount opportunities, cost of capital) are not provided, use a 6% cost of capital and note the assumption. Default automation efficiency gains to conservative benchmarks stated in Step 2.
Instructions
Step 1: Baseline the Current-State Cost
Establish the fully-loaded annual cost of running the process manually before any benefits are calculated.
- Direct labor cost: (Annual volume × minutes per transaction ÷ 60) × (fully-loaded hourly rate). Convert annual FTE cost to hourly using 2,080 working hours.
- Error/rework cost: (Annual volume × error rate) × (cost to resolve one error). If cost per error isn't given, estimate as (resolution minutes ÷ 60 × hourly rate) plus any downstream penalty.
- Cycle time cost placeholder: Note current cycle time in days — quantified as a benefit in Step 2, not a baseline cost.
- Present the baseline as a single Current Annual Process Cost figure with the labor and error components broken out.
Step 2: Quantify the Four Automation Benefit Levers
Apply each lever with a stated, conservative efficiency assumption. Always show the math.
Lever 1 — Labor Savings Automation removes or reduces manual handling time.
- Apply an automation rate: % of transactions that can be processed touchlessly (default: 60–75% for AP, 50–70% for reconciliations, 40–60% for close tasks — pick and state one).
- Labor savings = (baseline labor cost) × (automation rate) × (touchless efficiency factor, default 0.85 to account for exception handling).
- Express as annual dollars saved and FTE equivalents freed.
Lever 2 — Error Reduction Automation reduces exception and rework volume.
- Apply an error reduction rate (default: 50–70% reduction in error rate post-automation).
- Error savings = (baseline error cost) × (error reduction rate).
- Note both the hard cost savings and the risk/compliance benefit qualitatively.
Lever 3 — Cycle Time Compression Faster throughput frees capacity and improves downstream metrics.
- Estimate new cycle time (default: 40–60% reduction in days).
- Quantify the value: for close tasks, faster close = earlier decision data (state qualitatively + any headcount reallocation). For transactional processes, tie compression to Lever 4 where it unlocks working capital.
Lever 4 — Working Capital Improvement Faster, more accurate processing captures financial upside.
- Early payment discounts (AP): (annual spend eligible × discount rate captured × incremental capture % enabled by speed).
- DSO/DPO optimization: (days improved ÷ 365) × (relevant annual flow) × (cost of capital).
- If working capital levers don't apply to this process, state that explicitly and set the value to zero rather than inventing it.
Step 3: Total the Annual Benefit and Net It Against Cost
- Gross annual benefit = sum of all four levers.
- Net annual benefit = gross annual benefit − annual run/subscription cost.
- Flag any lever that dominates the case and note its sensitivity.
Step 4: Calculate ROI, Payback, and NPV
Model over a 3-year horizon.
- Payback period: implementation cost ÷ net annual benefit (in months).
- 3-year ROI %: ((cumulative 3-year net benefit − implementation cost) ÷ implementation cost) × 100.
- NPV: discount net annual benefits at the cost of capital (default 6%), subtract implementation cost. Show the year-by-year cash flow.
- Present all three headline figures together.
Step 5: Sensitivity Analysis
Test the case against pessimistic and optimistic scenarios.
- Build a 3-column table: Conservative / Base / Optimistic using ±30% swings on the two largest benefit levers and ±20% on implementation cost.
- Report payback period and 3-year ROI for each scenario.
- State the break-even condition: the minimum automation rate or benefit level at which the investment still pays back within the horizon.
Step 6: Recommendation and Assumptions Register
- Give a clear go / conditional-go / no-go recommendation with one-paragraph rationale.
- List every assumption used (efficiency rates, cost of capital, error costs) in a labeled register so the finance leader can challenge or replace them.
Output Format
- Structure: Sequential sections following the six steps, with a headline results summary at the top for executive scanning
- Length: 800–1,200 words plus tables
- Format: Structured document with narrative for reasoning and tables for the cash flow, sensitivity analysis, and assumptions register
- Sections: Executive Summary (headline ROI, payback, NPV) → Current-State Baseline → Four Benefit Levers → Net Annual Benefit → ROI / Payback / NPV → Sensitivity Analysis → Recommendation → Assumptions Register
- Tone: CFO-office analytical — precise, conservative, math-transparent; every number traceable to an input or a stated assumption
Quality Checklist
Before delivering output, verify:
- [ ] All four benefit levers are addressed — any lever set to zero has an explicit reason, not an omission
- [ ] Every calculated figure shows its formula or inputs — no unexplained numbers
- [ ] Efficiency and reduction rates are stated explicitly and fall within the benchmark ranges given
- [ ] Payback (months), 3-year ROI (%), and NPV ($) are all present in the executive summary
- [ ] Sensitivity table includes Conservative / Base / Optimistic with a stated break-even condition
- [ ] Assumptions register lists every rate, cost of capital, and estimated cost used in the model
Notes
- Conservative assumptions build credibility — a case that survives the pessimistic scenario is far more persuasive to a skeptical CFO than an aggressive base case
- Labor savings should be expressed in FTE equivalents as well as dollars; boards react to headcount reallocation differently than to abstract cost reduction
- If implementation cost is unknown, produce the model with a clearly-flagged estimate and show the break-even implementation cost — the point at which the investment stops paying back within the horizon
- Working capital upside is process-specific: it is significant for AP (early-payment discounts) but often negligible for internal close tasks — don't force a lever that doesn't fit the process
How to use this skill
This file is a Claude Code skill in Anthropic's standard format — markdown with YAML frontmatter at the top.
Installing in Claude Code:
~/.claude/skills/<folder-name>/SKILL.md— auto-loads in every session anywhere on your machine (recommended for skills you'll use across multiple projects)<your-project>/.claude/skills/<folder-name>/SKILL.md— auto-loads only when working in that project (recommended for project-specific skills)
Restart Claude Code or open a new session to register.
Using in Claude.ai, ChatGPT, Gemini, or other AI tools:
- Drag this file into a new chat, or copy and paste its contents
- Send your task or request as your next message — the AI will apply the skill where it fits
For Claude Code skill details: https://docs.anthropic.com/en/docs/claude-code/skills