Fixed-fee pricing is popular with Australian accounting and advisory firms for good reason. Clients like knowing the number up front, and a well-scoped engagement can be more profitable than hourly billing. The problem shows up later. The work quietly grows, the fee does not, and the margin you costed at proposal stage is gone by the time you lodge.
Scope creep is rarely one big event. It is a run of small yeses. An extra reconciliation here, a phone call that turns into an hour there, a director who emails three questions a week about a job you priced as a one-off. None of it feels worth a fee conversation on its own, so nobody has the conversation. Claude can help you see the creep while there is still time to act on it.
Why fixed-fee work quietly loses money
Most firms find out a fixed-fee job went over only when someone runs a work-in-progress report at month end, and by then the hours are sunk. The signals were there earlier, scattered across timesheets, emails, and your practice management system. They just never landed in one place where a partner could act.
Common signs a fixed-fee engagement is drifting include:
Recorded time on the job passing the hours you costed the fee against, with weeks still to run
A rising count of client emails or calls that sit outside the agreed deliverables
Repeated requests for the same information, usually because source records were incomplete
Add-on tasks that were never in the engagement letter but got done anyway to keep the client happy
The same job turning up in write-off discussions two years running
What scope creep detection actually means
You do not need a new billing system to catch creep. You need a regular, boring review that compares what you agreed to deliver against what the job is actually consuming, and flags the gaps early enough to matter. Claude suits this well because the inputs are messy and text-heavy: engagement letters, email threads, time narrations, and file notes. That is the kind of material a person can read but rarely has time to read across a whole client book.
The three questions Claude asks of every job
For each fixed-fee engagement, the review comes down to three questions.
What did we agree to do? Claude reads the engagement letter and lists the deliverables and any explicit exclusions.
What have we actually done? Claude reads time narrations and email activity and summarises the work performed, including tasks that were not in the letter.
Where is the gap, and is it worth a conversation? Claude flags jobs where out-of-scope work has passed a threshold you set, and drafts a short note the partner can use to raise it.
A simple monthly pricing review with Claude
A workable version of this runs once a month and takes a partner about half an hour to review, rather than the half day a manual pass would take. The steps look like this:
Export the open fixed-fee jobs from your practice management system, with budgeted hours and time recorded to date.
Point Claude at each engagement letter and the recorded time narrations for that job.
Ask Claude to rank the jobs by how far actual effort has run past the costed fee, and to explain in plain language what drove each overrun.
Have Claude draft a client note for the two or three jobs where a scope conversation is justified, matched to your firm's tone.
Review, edit, and decide. Nothing reaches a client without a partner signing off.
Because Claude drafts rather than sends, you keep full control of the client relationship. The tool does the reading and the first pass. The judgement about whether to absorb the overrun, reprice next year, or raise it now stays with you.
What this is worth in AUD
The maths is straightforward. Suppose a five-partner Sydney firm runs 300 fixed-fee jobs a year at an average fee of $4,500. If even 8 per cent of those jobs quietly run 15 hours over at a charge-out rate of $180 an hour, that is roughly 24 jobs losing about $2,700 in unbilled effort each, or around $65,000 a year in margin that never gets discussed. Catching half of it early, either by repricing at renewal or having the scope conversation before the work is done, is worth in the order of $30,000 a year to a firm that size. Running the review costs a fraction of that.
The bigger win is cultural. When creep is visible every month, partners stop treating write-offs as a fact of life and start treating them as a pricing decision they get to make on purpose.
Where to draw the line
This is a detection and drafting tool, not an autopilot for client pricing. Claude should never send a note, change a fee, or commit the firm to anything. Two guardrails keep it honest. First, a partner reviews every flag and every draft before a client sees it. Second, the thresholds that decide what counts as creep are set by the firm, not the model, so the review reflects your appetite rather than an arbitrary rule.
Client data also deserves care. Keep the review inside tools your firm already trusts for confidential records, and treat engagement letters and time narrations with the same discipline you apply to any client file under Australian privacy obligations.
Fixed-fee pricing only works when the fee and the effort stay in step. A short monthly review with Claude turns scope creep from an end-of-year surprise into a decision you make with time to spare. If you want help setting one up for your firm, book a short brainstorm and we will map it to how you already work.
