A small firm, with the cover of a big one.
AI for lawyers in firms of two to fifteen. The phone answered and intake taken, first drafts in your firm's style, a brief before every client meeting and the billing chase run. A lawyer reviews and signs everything.
The practice is small. The interruptions aren't.
A large firm has a receptionist, a bank of legal secretaries, a credit controller and a knowledge manager. A firm of six has a phone that rings while you're in court, a new client who called twice and went elsewhere, a precedent bank that lives in one partner's head, and time records that turn into invoices nobody chases.
So the lawyers do it themselves. Intake calls between meetings, engagement letters at nine at night, the same letter re-typed from last year's file, the overdue account left alone because the client is also a friend. None of it needs a practising certificate, and all of it lands on people who bill by the hour.
Most small-firm partners we talk to are brassed off with legal software that promised hours back and added another login. What they want is the routine work covered, by something that knows the firm.
AI for lawyers in a small firm: four jobs, covered.
A Foundation for a law firm usually starts with four parts of the AI workforce, each with one job, a boundary and a lawyer who reviews it.
Intake on the phone line
The AI receptionist answers every call the team can't, in the firm's name, day or night. It takes the caller's details, the other parties and the shape of the matter, checks the names against the firm's client records and flags any possible conflict for a lawyer, then books a consultation or passes the call on. It never gives advice and never tells a caller whether the firm can act.
Drafting in the firm's style
Trained on your precedents, your templates and the way your partners write. Engagement letters, client care letters, routine correspondence, file notes and first drafts of standard documents. The brain drafts; a lawyer edits and signs.
Matter briefs
Before each client meeting it pulls the matter: the last correspondence, key dates, what is outstanding and what the client asked last time, and lands a one-page brief an hour before. The top brass walk in prepared, and the file note is drafted afterwards.
The billing chase
Reads the practice management and accounting systems, turns recorded time into draft invoices for a partner to approve, and runs a polite, then firmer, chase on overdue fees. It never touches the trust account.
Privilege and privacy: what it never does.
Most of the work in a law firm can be prepared by someone else. The advice, the judgement and the signature can't. The install is built around that line.
It prepares, a lawyer decides
It answers the phone, takes the details, drafts and chases. It does not tell anyone what their legal position is, whether they have a case or what to do next. Anything that sounds like a request for advice goes straight to a lawyer.
Client files stay where they are
The install reads matters, documents and accounts in the systems the firm already runs, under the firm's own accounts, with access set by role the way your staff access is. Client material isn't pasted into public tools or used to train anyone else's model.
Nothing goes out unreviewed
No advice, letter or engagement leaves the firm without a lawyer approving it. The few messages it sends on its own, a booking confirmation or a payment reminder, follow rules the partners set, and every one shows up in the daily brief.
Your professional obligations stay exactly where they were. The install is designed to sit inside them.
Legal AI for a small firm, not a top-tier one.
Harvey, Spellbook and Lexis+ AI are serious products for legal research, contract review and drafting at scale, and they are mostly bought by larger firms with the budget, an IT team and a knowledge manager to roll them out. ChatGPT and Claude used directly are capable, but they don't know your precedents, your clients or your matters, and they don't answer your phone.
The Foundation is the other approach to AI for law firms of this size: a brain trained on your firm, wired into the systems you already use, with a workforce that does the routine work around the law rather than the law itself. If you add a research tool later, it sits alongside.
How it goes in.
It starts with the AI Strategy Intensive: two hours one-to-one with Titus, pre-work beforehand, then a written audit, a prioritised plan and a full proposal within three business days. It is $797, money-back, and credits in full to a build. If you'd like the wider view of what that diagnosis involves, our AI consulting page sets it out.
Then the Foundation: a three-week install of the AI brain, the data layer and the first AI workforce, owned by the firm from handover, code, IP and documentation included. After that it is kept polished, which is our name for AI training, optimisation, implementation and support as the practice changes. The full method is on the how we work page.
The pattern isn't new to us. Ian at AVIT Properties in London went from 80 hours of monthly reporting to two hours of review on the same data and drafting layers, Ray White Milldale runs Speed to Lead on the same first-response layer that answers a firm's phone, and we run our own firm on the system every day.
Questions law firms ask.
No. It takes the caller's details and the shape of the matter, books a time or passes the call to a lawyer, and says plainly that a lawyer will advise. What it can and can't say is written down with you before it answers a single call.
It checks the names a caller gives against the firm's client records and flags a possible match. It doesn't clear a conflict or decide whether the firm can act. A lawyer does that, with the flag and the call notes in front of them.
It stays in the systems the firm already runs, under the firm's own accounts. The install reads it where it lives, access follows the same roles as your staff, and client material isn't pasted into public tools. The detail is agreed in writing before anything is connected.
It reads the systems the firm already uses wherever they allow access, which most current practice management and accounting systems do. Which systems, and what it reads from each, gets confirmed during the Intensive before any build is quoted.
Both. It is built in Auckland and installed across New Zealand, and Australian firms are installed remotely. The professional rules differ across the Tasman, and the boundary is set to your side of it.
The first step is the AI Strategy Intensive: $797, two hours one-to-one with Titus, money-back, and credited in full to a build. The build is scoped during the Intensive, so it fits the firm rather than a tier.
Other places a small firm loses time.
Spend the hours on the law.
Thirty minutes with Titus. We look at where the routine work lands in your firm, and whether a Foundation install fits.
Book a discovery callNot ready for a call? Take the three-minute Business Stress Test.