Does AI kill the billable hour?

No, but it makes it indefensible for a shrinking category of work. The international evidence points to bifurcation rather than collapse. Repeatable, precedent-based matters move toward fixed and value-based fees because AI has made them fast and predictable. Contested, bespoke, high-judgment work stays hourly because the time genuinely is the risk. The billable hour is losing territory, not disappearing.

Why do lawyers bill in six-minute increments?

Six minutes is a tenth of an hour, so it lets a firm convert time into a fee with simple arithmetic. It became standard because it was the most practical way to measure and recover professional effort before software could track work automatically. The unit is an accounting convenience, not a measure of what the work was worth to the client.

What does the New Zealand rule actually say about legal fees?

Rule 9.1 of the Lawyers and Conveyancers Act (Lawyers: Conduct and Client Care) Rules 2008 sets out thirteen factors relevant to whether a fee is fair and reasonable. Time and labour expended is one of them, listed at (a). The others include complexity, urgency, skill, the value of the matter, and whether the fee was fixed or conditional. Any specific question about fees is one for the firm and its own advisers.

What are alternative fee arrangements?

Any pricing structure that is not purely time multiplied by rate. The common ones are fixed fees per matter type, capped fees, retainers or subscriptions for ongoing advice, and success-based components tied to an outcome. New Zealand firms already use fixed fees widely for conveyancing, wills and most immigration work, so the structure is familiar rather than novel.

What happens to a firm's revenue when AI makes the work faster?

Under hourly billing it falls, unless something else changes. Clio's 2026 research on solo and small firms found 71 percent of solos and 75 percent of small firms using AI, but only 32 percent and 31 percent respectively had seen it lift revenue, and 86 percent and 78 percent had made no pricing changes at all. Speed without repricing transfers the gain to the client.

NZ law firm pricing models and AI

Charge for the outcome, not the hour. If AI takes a fixed-scope matter from 90 minutes to 15, an hourly firm has just cut its own fee by 83% for identical work. Clio's 2026 research found most small firms that adopted AI changed nothing about their pricing, and most saw no revenue growth from it.

That is the whole problem in three sentences, and it is the one nobody selling AI into law firms wants to open with.

The unit of account is the problem

New Zealand lawyers bill in six-minute increments, at somewhere between $250 and $800 an hour depending on the work and the firm, with the average sitting around $350 to $400. Six minutes is a tenth of an hour. It exists because it makes the arithmetic easy, not because anyone believes six minutes is a meaningful unit of legal value.

For most of the profession's history that did not matter much. Time was a decent proxy for effort, and effort was a decent proxy for value. The proxy held.

AI breaks the proxy. Not everywhere, and that distinction matters, but it breaks it precisely where the work is repeatable and precedent-based, which is a large slice of what a general practice actually does day to day.

The uncomfortable arithmetic

Take a fixed-scope matter that used to take 90 minutes of a lawyer's time. At $400 an hour, that is $600 of billable work.

Now assume AI-assisted drafting and a decent template library take the same matter to 15 minutes of lawyer time, with the output reviewed and signed off to the same standard. At the same rate, the firm bills $100.

Same client. Same document. Same professional responsibility for the result. Eighty-three percent less revenue.

(That example is illustrative. No source in this space publishes a task-level worked example in New Zealand dollars, which is itself telling, so treat the shape of the arithmetic as the point rather than the specific numbers.)

Every firm that has adopted AI has run some version of this calculation, usually privately, and usually with a slightly sick feeling. It is why "we're looking into it" is such a common answer. The efficiency is real. The efficiency is also, under hourly billing, a direct cut to the top line.

The data says firms are not resolving it

Clio's 2026 Legal Trends for Solo and Small Law Firms is the clearest read available. Adoption is high: 71% of solo practitioners and 75% of small firms report using AI. The operational benefits are real and reported consistently, with 64% saying work quality improved and 60% responding to clients faster.

Then the numbers turn.

Only 32% of solos and 31% of small firms have seen AI lift revenue, against 59% at enterprise firms. And 86% of solo firms and 78% of small firms have made no pricing changes at all since adopting it.

Read those two findings together and the mechanism is obvious. Firms got faster and kept charging by the hour. The gain went to the client, for free, without anyone deciding that it should.

Worth noting the source: Clio sells practice management software, so it is not a disinterested party in a finding about pricing discipline. The figures still line up with what the argument predicts, and nothing in the counter-literature disputes them.

The answer the market gives is not an answer

Read the New Zealand implementers writing about AI for law firms and you will find the problem named accurately in the opening paragraphs. Margin pressure. Hourly fees under strain. Then watch what happens to it.

The resolution offered is almost always one of two things. The first is recovery: AI captures time that was worked but never logged, so the firm bills more of what it already did. The second is capacity: the same lawyers handle 20% to 30% more matters at the same rates.

Both are real. Both are worth having. Neither answers the question.

Recovery only helps to the extent a firm was leaking billable time, and it is a one-off correction rather than a structural fix. Capacity only converts to revenue if there is more work waiting, which for a firm in a small regional market is an assumption, not a given.

More to the point, neither says the quiet part. A firm that completes a fixed-scope matter in a fraction of the time and charges the same fee has not recovered anything. It has kept a fee that no longer matches the work. That is a defensible commercial position, and it may well be the right one, but it is a pricing decision and it should be made deliberately rather than by leaving the timer running.

The honest counter-case

The billable hour is not dying, and any piece telling you it is should be read carefully.

Wolters Kluwer's 2026 Future Ready Lawyer survey found over half of respondents expect AI to reduce billable hours and push work toward alternative fee arrangements. Its own framing is that the billable hour is unlikely to disappear and that its role is changing. Thomson Reuters' coverage of the ILTACON 2025 discussion reached the same conclusion: hybrid pricing accelerates, hourly billing survives for uncertain and high-judgment work, and nothing gets fully replaced inside five years. Bloomberg Law has been blunter still, calling it the most resilient pricing model there is.

Both of those surveys are international, and the Wolters Kluwer sample contains no New Zealand or Australian respondents, so read them as direction rather than local measurement.

The accurate version is bifurcation. Hourly billing is becoming indefensible for repeatable, AI-compressible work, and remains entirely legitimate for contested, bespoke work where the time genuinely is the risk being priced. A litigator cannot know in advance how many hours a defended hearing will take, and pretending otherwise helps nobody.

New Zealand already has the structure

Here is the part that makes this practical rather than theoretical. New Zealand firms do not need persuading that fixed fees work, because they already use them.

Conveyancing runs fixed, commonly $1,200 to $2,500 for a house purchase. Wills run fixed, roughly $300 to $800. Most immigration work is quoted as a fixed fee per application. Meanwhile litigation, contested family law and criminal defence stay hourly. The profession sorted its work into predictable and unpredictable a long time ago and priced each accordingly.

That existing line is the answer. It just needs to move.

The matters AI compresses hardest are precedent-based drafting, intake, conflict checking and standard-form documents, which is almost exactly the profile of work that already sits on the fixed-fee side. The question for a practice in 2026 is not whether to adopt an unfamiliar pricing model. It is which matter types have quietly crossed from unpredictable to predictable, and whether the fee structure has been updated to match.

And the rules already accommodate this. Rule 9.1 of the Lawyers and Conveyancers Act (Lawyers: Conduct and Client Care) Rules 2008 sets out thirteen factors bearing on whether a fee is fair and reasonable. Time and labour expended is one of them, listed first at (a). Complexity, urgency, skill, the value of the matter and whether the fee was fixed or conditional are among the others. Time is one factor of thirteen. The framework was never as time-bound as the habit.

What saved hours are actually worth

The value of an hour you no longer have to work depends entirely on how you charge for it.

We built a practice management platform for My Wellness, a health practice, where the automation freed up more than ten admin hours per practitioner every week. That converted cleanly into value, because the practice bills per appointment. Hours saved became appointments available. The saving and the revenue pointed the same direction.

/ From the workshop · Health platform · Queenstown

My Wellness: 10+ admin hours saved per practitioner every week.

Read the case study →

Run the identical build inside a firm billing in six-minute units and the arithmetic inverts. Same technology, same hours released, opposite effect on the top line. This is why "how many hours will it save us" is the wrong first question for a law firm, and why an efficiency pitch tends to land badly with partners who have already worked out what efficiency does to their revenue.

The better first question is which matters could be priced differently if they became reliably fast.

Why now rather than later

New Zealand firms are having a good year. Thomson Reuters' 2026 Report on the State of the New Zealand Legal Market shows revenue up more than 10%, demand up more than 5%, and the first double-digit profit growth since 2021, with average margins around 43%.

None of that is attributable to AI. The report puts it down to a recovery in transactional demand after a weak 2022 and 2023. But it does mean the conversation can be had from a position of strength, which is a considerably better time to restructure how you price than the year you are forced to.

The firms that will handle this well are the ones auditing their matter types now: which are genuinely repeatable, which are genuinely bespoke, which have moved from one column to the other in the last eighteen months without anyone updating the engagement letter.

That audit is not a technology project. It is a pricing exercise that happens to need someone who understands what the technology can actually compress.

If you are working through which parts of your practice have quietly become predictable, a short call is the fastest way to find out whether it is worth doing properly.

If AI halves the work, what should a law firm charge?