What CPA Firms Need Before They Move Off the Billable Hour

The pricing argument inside the profession is finished, and most firms haven't noticed they won it. Ignition's 2025 US accounting and tax pricing benchmark, published July 2025 from 219 US firms, puts hourly billing in the minority in every service line it tracks — 3% for tax preparation, about 4% for bookkeeping, 10% for CFO and controller work, and 17% for tax planning and advisory. Fixed fee runs the book. The question is no longer whether to leave the billable hour.

The question is what number you write down when you do.

I sell software into this problem, so weigh the rest accordingly. The observation that keeps coming up on calls is not about pricing philosophy. Firms have been sold five of those. It is that when partners sit down to set a fixed fee for next year, the only number in front of them is last year's invoice, and they add a percentage to it.

The hours are being pulled out from under the invoice

Thomson Reuters Institute's 2026 AI in Professional Services report has 40% of professionals saying their organizations now use generative AI, up from 22% a year earlier — nearly double. Bloomberg Tax, on March 5, 2026, quoted PwC's US tax leader Krishnan Chandrasekhar saying time is "becoming less and less of relevance," and cited tax advisory work compressing from 20 hours to 10. RSM's Sergio de la Fe framed the whole problem in one question in that same piece: is the saving margin for the firm, or a price cut for the client.

Clients are answering it for you. Accounting Today, March 24, 2026, reported a General Assembly survey of 258 director-level-and-up leaders at US and UK firms of 1,000 or more employees, so read it as the leading edge rather than a portrait of the mid-size firm: 79% of professional-services leaders said AI is changing pricing conversations, 35% of the accounting respondents said clients are questioning their pricing because of it, and 38% said their firm was getting ahead of it. If it's reaching the thousand-person firms first, it is not stopping there.

An hour saved is an hour you can't bill. Every efficiency investment a firm makes right now is an investment in shrinking its own invoice, and that math only stops being hostile when the invoice stops being made of hours.

The realization report is a post-mortem

The claim that firms have no visibility into WIP and realization is wrong, and any partner running CCH Axcess Practice will correct you on it. The platform ships Client Billing Realization, WIP/AR Reconciliation, and Staff Realization and Productivity. The data exists.

It is the wrong shape for pricing. It is aggregated by partner and by staff, it is cut for the period rather than the engagement, and it arrives weeks after the decisions that produced it. It tells you realization was 87% last quarter. It does not tell you, on the day you release this bill, that this engagement has run over scope three years running and that the drift sits in three specific line items.

That is the difference between a report and an instrument, and it is the whole gap.

Reconstruct twenty-four months before you touch a fee

Do this before any client-facing change, and do not let it become a project with a steering committee.

Pull two years of engagements and compute, for each one: hours at standard cost rather than billing rate, realization, write-ups and write-downs with a reason attached, and hours that fell outside the engagement letter. Then build the variance distribution. You want the list of engagements that ran 20% or more over estimate and the reason each one did: scope, client responsiveness, staffing mix, or review churn. Four buckets is enough.

Then segment the book into profitable, marginal, and losing. Firms that run this honestly usually find a long tail of individual returns underwater and one or two large, well-known clients that the firm has been proud of for years while losing money on them.

The 2025 Rosenberg Survey has income per equity partner up 3.2%, to $615,000, on revenue growth of 7.9%. Revenue is growing and not converting — and the mid-size firm, without the scale to absorb a soft engagement or the leverage of the largest firms, has the least slack when it doesn't.

Keep the timesheet. Stop invoicing from it.

Ron Baker has been right about price-led costing for thirty years and has only a small minority of the profession to show for it, a number that has barely moved. The plank that costs him the rest is timesheet abolition, because partner compensation, firm valuation, and M&A multiples are all still computed off rates, realization, and leverage. Asking partners to change how clients are charged and how partners are paid in the same year is asking for one thing too many.

Keep time entry. Redefine it, in writing, as internal cost accounting and capacity planning, and say plainly that it will never again be the basis of a client invoice. Announce that before anyone asks. Staff who are not told will assume the firm is going back to hourly by stealth and will start padding accordingly, which corrupts the exact dataset the whole exercise depends on.

Sequence the repricing by realization, not by relationship

The instinct is to start with the best clients, because those conversations are pleasant. Start with the worst realization instead. Phase 0 already named them.

Price from cost to deliver plus target margin plus a value adjustment, and write standard scopes first: what is in, what is explicitly out, what the client owes you and by when, and the trigger that converts extra work into a signed change order before the work starts. Pick a threshold and hold it. Anything out of scope over two hours or $500 needs a change order, with a named owner on every engagement. Fixed fee without that discipline is a permanent unfunded write-off, and it is the most common way this transition fails.

Ignition's 2026/27 benchmark, from roughly 350 US firms, has 77% planning fee increases in the next 6 to 12 months and more than 60% reporting little or no client resistance after raising prices. Ignition notes its year-over-year comparisons are directional given sample differences, and they are a vendor with a product in this space. Read the direction, not the decimal.

Do not hand the AI savings back as a discount

Clients are going to ask. The answer is not a lower fee. It is more scope inside the same fee and a turnaround-time commitment you can actually hold, which is worth more to a client than four percent off and costs you less. A firm that reflexively passes efficiency through as price reduction has automated its way to a smaller business.

What replaces realization on the scoreboard

Gross margin per engagement first. Then effective hourly rate — collected fee divided by actual hours — the bridge metric that lets you have this conversation in language partners already speak while measuring something new. Then revenue per FTE. Then scope-variance rate: the share of engagements that exceeded scope, and the share of that overage you recovered through a change order.

Rebuilding compensation around those numbers is the phase most firms skip, and skipping it is why the pricing change reverts within two seasons.

None of this requires new software. It requires the engagement-level history to exist in a form somebody can act on, and for most firms that history is currently spread across pre-bills, adjustment fields, and the memory of whoever ran the bill.

That history is what Billswarm's WIP software produces as a byproduct of clearing bills: line-item WIP, coded write-downs, and per-engagement realization, generated where the billing decision is made instead of a quarter later. It deploys inside the firm's own infrastructure, because a firm's engagement economics is the most sensitive dataset in the practice and there is no good reason for it to sit anywhere else. Questions go to austin.eisenman@billswarm.com.