Five things the profession believes about AI and small firms. The data disagrees.
Most of the NSW profession is sole and small. Most of what gets said about AI and small firms is wrong. Here is what the Australian numbers actually show.
Eighty-two per cent of NSW private practices are sole practitioners. Almost nothing written about legal AI is written for them, and much of what is said about them does not survive contact with the data. Five myths, checked against the Australian evidence.
Small firms are the innovators, so they will lead on AI
The opposite is happening
Solo and small firms led the last technology wave. They moved to the cloud and to practice-management software faster than the big firms, and the profession got used to thinking of them as the nimble ones. On AI that has reversed. LEAP's 2026 survey found 16 per cent of Australian respondents using legal-specific AI daily, against 49 per cent globally, with small firms trailing large ones. The group that was early last time is late this time. The interesting question is why, and the answer is in the next four myths.
Lawyers who avoid AI just do not understand it
Their caution is professional judgment
The convenient story is that the holdouts are technophobes who will be educated out of it. The Australian profile says otherwise. Just over half of the country's sole practitioners were admitted 15 years or more ago. This is an experienced cohort, and what it understands better than anyone is what a fabricated authority does to a matter and what verification actually costs. LEAP recorded 32 per cent of Australian respondents reporting low or no trust in AI providers, the highest of any market it surveyed. That is not ignorance. It is people who have read the cases pricing the risk correctly.
The scepticism is not a knowledge gap. It is experienced lawyers pricing the risk correctly.
AI solves the small firm's biggest problem: capacity
It solves the problem you do not have
The whole industry pitch is "do more work, faster." But for most sole practitioners the binding constraint was never the volume of work they could process. It was the volume of clients walking in the door. Clio's research captures it in a single line from a solo practitioner: if they could find the clients, they could do ten times the work. AI clears the desk faster and then leaves you staring at the actual bottleneck, which is acquisition. A tool that drafts your advice in half the time does nothing for an empty diary. The firms getting real value have worked out that efficiency is only half the machine.
The big firms carry the real AI risk
The risk concentrates at the small end
A global firm putting AI into client work has an IT department, a risk committee, vetted enterprise tools and layers of review. The sole practitioner has none of that, and is the most likely to reach for a free public chatbot, because free is what a one-person business reaches for. The Law Society has been blunt: entering confidential client information into a public tool is akin to putting it in the public domain, with privilege potentially waived. The exposure is identical in kind and worse in practice for the small firm, because there is no second set of eyes. The Australian referrals bear it out. They have come from family law, immigration and general litigation, the work of small practices, and in each the failure was the same, an output filed without an independent check.
The numbers prove AI pays off for small firms
For this segment, the numbers barely exist
Hours saved, revenue lifted, matters cleared: the figures are everywhere and almost all of them are vendor surveys, North American data, or studies that blend Australia with New Zealand and mix firm sizes together. There is no independent, Australia-specific dataset isolating what AI is actually doing to a sole practitioner's revenue, capacity or client numbers. What is documented is narrower and more useful: the work is shifting toward fixed and capped fees, which sits awkwardly with billing by the hour for work AI did in minutes, a tension the College of Law has flagged against the duty that costs be fair and proportionate. Beyond that, certainty is being sold, not measured.
The throughline across all five is the same. The small-firm position on AI is not a failure to keep up. It is a different and largely rational response to a technology whose loudest claims were written for someone else's practice. The sole practitioner who reads the hype and feels behind has it backwards. The useful move is not to adopt faster than the marketing says. It is to be clear-eyed about which of these five myths the next sales call is built on.