Agencies that bill by the hour have a quiet problem with Claude Code. A tool that helps your developers ship the same work in a third of the time sounds like a win, right up until you notice the invoice shrinks with the hours. If a build that used to take 40 hours now takes 14, and you charge $180 an hour, you have just handed the client a $4,680 discount for doing better work. That is the pricing problem, and pretending it does not exist is the most expensive option on the table.
Where the money leaks
Hourly billing ties your revenue to effort, not to the result the client actually bought. That was a workable proxy while effort and outcome moved together. Claude Code breaks the link. When it drafts the boilerplate, writes the first pass of tests, explains an unfamiliar codebase, and handles the routine refactors, the hours fall away while the outcome stays the same or improves. Your most capable developers become the ones who bill the least, which is exactly backwards.
For a Sydney studio running a team of six, the leak shows up in a few predictable places:
Fixed-scope work quoted in hours. You estimated 60 hours, delivered in 22, and either eat the gap or awkwardly pad the timesheet.
Retainers measured in capacity. A 40-hour monthly retainer now covers far more output, so the client feels the value climb while your revenue sits flat.
Change requests. The small tweaks that used to justify a few billable hours each now take minutes, and clients notice.
Junior staffing math. Work you once staffed with three juniors is done well by one mid-level developer and Claude, collapsing the margin you built into blended rates.
Four ways Australian agencies are repricing
None of these are new inventions. They are the models professional services firms already use when effort stops being a fair measure of value. Claude Code just makes the move urgent rather than optional.
Value-based project pricing. Quote the outcome and the business case behind it, not the hours. A booking system that saves a client $120K a year in manual admin is worth a fixed fee that reflects that number, whether it takes you 14 hours or 40.
Productised fixed-scope offers. Package the work you do repeatedly into named deliverables with a set price and a set turnaround. Faster delivery becomes your margin, not the client's discount.
Outcome retainers. Replace the capacity retainer with one tied to what you maintain and improve each month: uptime, features shipped, tickets closed. The client buys progress, not a bucket of hours.
Blended advisory plus build. Charge for the judgement, architecture, and review that Claude does not replace, and treat the mechanical build as the cheap part it has become.
A worked example
Say a Melbourne agency wins a customer portal build. Under the old model they quote 160 hours at $185, for $29,600, and make a healthy margin at 160 hours of real work. With Claude Code in the workflow the same portal takes the team 70 hours. Bill it hourly and the project is now worth $12,950, a drop of more than half, for the same shipped product.
Reprice it as a fixed-fee deliverable at $27,500 and the picture changes. The client pays a little less than the old quote, feels good about it, and gets the portal weeks earlier. The agency does 70 hours of work for $27,500, an effective rate near $390 an hour, and the freed capacity goes to a second client. The tool created the surplus. The pricing model decides who keeps it.
The awkward client conversation
The objection you will hear is that fixed fees feel like a black box next to a timesheet. It is a fair concern, and the answer is not to hide the efficiency. Be plain about it. Clients do not buy hours because they enjoy hours; they buy a working outcome by a date they can trust. A fixed price with a clear scope and a firm delivery date gives them more certainty than an open-ended hourly estimate ever did. The agencies handling this well in Australia are the ones treating the shift as a selling point, not a secret.
There is also a governance angle worth naming for clients in regulated sectors. If you are shipping software that touches personal data under the Privacy Act, the review and accountability you provide around Claude's output is part of what they are paying for, and it belongs in the price. That work does not shrink when the typing gets faster.
What to do this quarter
You do not need to rip up every contract at once. Start by measuring the gap: for your last five projects, compare quoted hours against actual hours now that Claude Code is in the mix. That number is your current discount, and it is usually larger than people expect. Then pick one repeatable offer and reprice it as a fixed deliverable before your next proposal goes out. Keep hourly billing for genuinely open-ended discovery work where the scope really is unknown, and move everything with a predictable shape onto value-based pricing.
The agencies that win the next few years will be the ones who let the tool make them faster and kept the upside for themselves. If you want a second pair of eyes on how to reprice your services around Claude, we help Australian teams do exactly this. Book a brainstorm and we will work through the numbers with you.
