Close Health Score
Self-reported ranges produce a benchmarked score and an estimated leak. Band width reflects uncertainty, never false precision. Directional only — never a verified figure.
The method, in full
From a free score to a proven, deployed automation
Seven stages, each producing something concrete — a benchmarked score, a written finding, a working AI finance skill, a live automation and measured savings. This is the sequence a FlyCFO engagement actually runs through.
One measured path
Seven stages, seven decision gates
Score, Diagnose, Findings, Skill, Automate, Verify, Operate. Each stage below opens onto what it actually produces — not just a name on a rail.
- ScoreClose Health Score
- DiagnoseMap manual steps
- FindingsBoard-ready report
- SkillCited finance skill
- AutomateDeploy one workflow
- VerifyMeasure the savings
- OperateSarah stays on
What each stage actually produces
The method, stage by stage
01 · Score
Score
The free Close Health Score benchmarks your month-end close from the ranges you report and estimates where cash and hours leak to manual work. No logins, no ledger data.
- A benchmarked score (0–100) with a healthy / attention / critical band
- An estimated annual leak to manual finance work
- Your first automation targets, ranked by likely payback
Decision gate — is there a clear enough payback to justify a paid assessment, or are you already lean enough to wait?
02 · Diagnose
Diagnose
Inside the paid assessment, we map how the close actually runs — the manual steps, reconciliations, journals and hand-offs — under an agreed, least-privilege access model.
- A step-by-step map of the close, with manual effort quantified
- A systems and data-flow picture across ERP, AP, payroll, cards and banks
- A shortlist of automation candidates, each bounded and testable
Decision gate — which single automation has the highest payback and lowest risk to deploy first?
03 · Findings
Findings
The diagnosis becomes a written, board-ready findings report — in language your board and audit committee can act on, with every number graded by evidence.
- Board-ready findings, ranked by payback and risk
- Evidence grade against each quantified claim
- A recommended sequence — including "don't automate this yet" where true
Decision gate — do the findings justify building, or is the honest recommendation to hold?
04 · Skill
Skill
We build a working AI finance skill grounded in your own numbers, policies and definitions — one that answers finance questions and drafts work with citations back to the source.
- A finance AI skill scoped to your chart of accounts and policies
- Citations on every answer, so nothing is an unsourced guess
- A skill you keep, whether or not you proceed to a larger build
Decision gate — does the skill answer real finance questions accurately and with sources before it's relied on?
05 · Automate
Automate
We deploy the highest-payback automation — not a prototype, a live, managed workflow with monitoring and an agreed recovery path.
- One deployed automation running against your real close
- Monitoring, alerting and a documented recovery procedure
- Ownership and access recorded, so nothing depends on tribal knowledge
Decision gate — has the automation passed its acceptance checks before it's trusted in the close?
06 · Verify
Verify
We measure what the automation actually returns — hours and dollars — rather than projecting a business case and hoping.
- Measured hours and cash returned, compared to the pre-automation baseline
- An evidence-graded savings record you can take to the board
- A decision on whether to extend into a larger automation build
Decision gate — are the verified savings enough to justify automating the next step?
07 · Operate
Operate
Live automations need an owner. FlyCFO Managed keeps them healthy and improves the close every month — Sarah, your AI finance operator, stays on.
- Ongoing monitoring, failure recovery and change handling
- A monthly service report and reserved improvement capacity
- A clear boundary between platform operation and finance policy ownership
Decision gate — is there a named owner and a healthy operating cadence going forward?
How confidence is built
Three evidence levels, three kinds of confidence
A number is only as honest as the evidence behind it. FlyCFO names the level explicitly rather than presenting every figure with equal confidence.
Assessment findings
Working with your actual data, we quantify manual effort and cost per step and grade each claim. Enough to prioritise and to design the first automation.
Verified savings
After the automation is live, we measure the hours and dollars it actually returns against the pre-automation baseline. The strongest evidence we publish.
From estimate to evidence
Estimate → Diagnose → Deploy → Verify
We never ask you to trust an estimate as if it were verified. The sequence turns a directional score into measured savings, one honest step at a time.
01 · Estimate
Close Health Score
A benchmarked, directional score from your self-reported ranges. Useful for direction, never verified.
02 · Diagnose
Findings & skill
We work with real data to quantify effort and build a cited finance skill.
03 · Deploy
One automation
The highest-payback workflow goes live, monitored and managed.
04 · Verify
Measured savings
We measure the hours and cash actually returned before any larger build.
Shared accountability
Who owns what
FlyCFO owns
- The Close Health Score model and the assessment findings
- Building and grounding the AI finance skill
- Designing, deploying and testing the first automation
- Monitoring and recovery for what we deploy
- Measuring and reporting verified savings
You own
- Naming the true owner of each finance process
- Providing agreed, least-privilege system access for the assessment
- Confirming finance policy, definitions and accounting treatment
- Approving which automation is deployed first
- Finance policy and controls decisions going forward
Illustrative
An example assessment timeline
For a mid-sized finance team running one close cycle. Actual duration depends on system access, data readiness and how quickly decisions can be made.
Score & kickoff
Close Health Score, then agree scope, access and the close cycle to observe.
Diagnose
Map the close, quantify manual effort and shortlist automation candidates.
Findings & skill
Board-ready findings delivered; the cited AI finance skill built and reviewed.
Automate
Deploy and test the first automation against the live close.
Verify
Measure the hours and cash returned; decide whether to extend.
Illustrative timeline only — not a delivery commitment for any specific engagement.
Questions, answered
The things people ask about the method
Do I have to do the paid assessment to get value?
No. The Close Health Score is free and useful on its own — a benchmark and a directional cost estimate. The paid assessment is for when you want verified numbers, a working skill and a deployed automation.
Is the Close Health Score a guarantee?
No. It's a directional estimate from deterministic rules on self-reported ranges. Verified figures come from the assessment working with your real data.
What if the findings say we shouldn't automate yet?
Then that's what we'll tell you, with the reasoning and the conditions that would change it. You still keep the findings and the working finance skill. We're designed to give an honest read, not to force a build.
Do you need access to our systems?
Not for the score. For the assessment, yes — under a least-privilege, agreed and revocable access model that's documented before anything is connected. Never shared passwords.
Start with a number
Score your close before you commit to a build.
Get a benchmarked Close Health Score now. Turn it into an assessment when the payback is clear.