DIY, Buy or Build: The Capacity Math on AI for a Fractional CFO Firm

A fill-in-your-numbers worksheet, built on published fee and benchmark data, for deciding between DIY AI, built-in tool AI and a built workflow.

Should a fractional CFO firm do AI drafting by hand in a chat tool, rely on its reporting tool's built-in AI, or pay to have a workflow built? The honest answer depends on three numbers you already know: how many clients you carry, how you bill them, and how many hours the monthly narrative takes. This post walks through the math with published figures, and shows where it doesn't add up.

Start with how you bill

This matters more than anything else.

The AICPA and CPA.com 2024 benchmark survey of client advisory services (CAS) practices, covering 206 practices and calendar 2023 data, found only 10% bill primarily by the hour (CPA.com, Dec 2024). Most charge fixed recurring fees. For fractional CFOs specifically, published retainers run about $3,000 to $10,000 a month according to GoFractional (2026), and $3,000 to $12,000 a month with hourly rates of $150 to $500 according to Eightx (Jun 2026), a firm that sells these services.

If you bill a fixed monthly fee, every hour you save drafting the narrative becomes margin, or room for another client.
If you bill by the hour, faster drafting can mean a smaller invoice. The case for AI then rests on quality, capacity or keeping clients, not on hours saved. Jason Hope, the fractional CFO profiled in CFO Dive (Aug 2026), bills hourly and still reports big gains, so it can work, but the math is different.

Why capacity is the number to watch

The same CPA.com survey found that CAS practices which keep investing in technology serve about 100 clients, against an average of 67, and that net client fees per professional were $156,250 (CPA.com, Dec 2024). The survey shows correlation, not cause: firms that invest in technology may differ in other ways. But it is the best published benchmark we found, and it points at the right question: how many clients can each CFO carry well?

Demand looks healthy overall: CFO Dive reports a 2026 BTG talent report finding interim CFO demand up 14% year over year (CFO Dive, Aug 2026). But in practitioner forums, finding new clients is one of the most common complaints fractional CFOs raise. If AI frees capacity, you need a plan for filling it, or the saving is only an easier month.

A worked example (arithmetic, not a forecast)

Take a firm with 10 clients on a $5,000 monthly retainer, the middle of the published ranges. That's $50,000 a month in fees.

Now fill in your own numbers:

Input	Your figure
Clients with a monthly narrative	[A]
Hours per narrative today, writing and rewriting	[B]
Hours per narrative with a reviewed AI draft (measure it over two closes)	[C]
Hours saved per month	[A] x ([B] - [C])
Hours a new client needs per month	[D]
Extra clients that time could support	([A] x ([B] - [C])) / [D]

Don't guess [C]. Vendor estimates of time saved on board packs are mostly unmeasured, and one published estimate we read was internally inconsistent. Run the narrative template on two clients for two closes and time the edits.

If the result is less than one extra client's worth of hours, AI is a quality and stress improvement, not a growth plan. That can still be worth it.

What each option costs

Do it yourself in Claude or ChatGPT. A business account (so client data isn't used for training by default; see our data rules post) plus staff time to copy data in, run the template and keep memory files up to date. Rework lists Claude Team at $25 to $125 per seat per month with a five-seat minimum (Rework, Jul 2026; check the vendor for current pricing). The hidden cost is consistency: each CFO does it a little differently.

Use your reporting tool's AI. If you're on Fathom Pro ($450 a month for 25 companies, plus $17 per extra company, per Fathom's pricing page, fetched Oct 2026), the Commentary Writer is included, so the marginal cost is zero. Jirav partner pricing runs about $50 to $150 per client per month (Rework, Jul 2026), which tells you what firms already accept paying per client for tooling.

Have a workflow built. A one-time cost to connect the ledger and forecast, set up per-client memory and a review step, then a smaller monthly cost to keep it working as clients and charts of accounts change. Worth comparing against the analyst hours it replaces, not against a software subscription.

A simple decision rule

Under about five clients with a monthly narrative: do it yourself with the template. The setup cost of anything else won't pay back.
Built-in AI already good enough on your test client: use it, keep a memory file alongside, done.
Many clients, fixed fees, and the narrative still takes most of a day each: a built workflow starts to make sense, especially if the drafts need context your reporting tool can't see.
Billing hourly: start with quality and capacity, not hours saved, and talk to clients about value pricing before you invest.

Three numbers to track for two months

Edit minutes per narrative after the AI draft.
Share of the draft that survives review.
Clients per CFO. The one that ultimately matters.

If the first two improve and the third doesn't move, you have an easier month. If the third moves, you have a bigger practice. The CPA.com survey also found that niche-focused CAS practices had 38% higher median CAS revenue (CPA.com, Dec 2024), which is worth remembering when you decide what kind of clients to fill the new capacity with.

For the full picture of tools and workflows, start with how fractional CFOs use AI and what each reporting tool's AI covers.

Want a second pair of eyes on your monthly cycle? Lunan offers a free conversation: we ask how your close-to-board-pack month runs, which tools you use and how you bill, then send a short written note on where drafting would save the most time and where your current tools already cover it. Start a free conversation.