A 12-partner commercial law firm

The firm was not short of demand. It was short of a way to answer demand when it arrived. Commercial clients research at night, compare two or three firms in an evening, and instruct the one that responds with confidence first. A form that sat unread until the morning was a form that had already lost.

Industry figures put it bluntly: in one audit of 433 B2B companies, 55% took more than five working days to reply to an enquiry, or never replied at all (Drift Lead Response Report, 2017). For a commercial practice where a single instruction can be worth five figures, that is not a service problem. It is a revenue leak with a number attached to it.

Where the work was leaking

We tracked a month of enquiries before touching anything. Just under half arrived outside office hours. Of those, the firm replied to most within a working day, which felt responsive internally and read as slow to a prospect who had already booked a call elsewhere.

The pattern was consistent. An enquiry would land at 8pm on a Tuesday. The fee-earner would see it at 9am on Wednesday and reply by lunch. By then the prospect had already spoken to two other firms and instructed one of them. The work was lost in the gap, not in the pitch.

What we built

The site came first, because a commercial client decides whether a firm is serious in the first few seconds. Then the AI receptionist: not a chatbot answering FAQs, but a worker briefed on the firm's practice areas, fee structure, and the questions a partner needs answered before a first meeting.

• Matter type and urgency confirmed in the first exchange.
• Budget and scope established before a partner's time was committed.
• A conflict-check prompt and a drafted matter note handed to the partner with the booking.
• Three weeks of professional follow-up on any enquiry that went quiet.

Why it reads as the firm, not a bot

The objection partners raise first is tone: a commercial client will not tolerate being handled by something that sounds automated. So the AI employee was briefed in the firm's own register, given the boundaries of what it could and could not commit to, and set to escalate anything sensitive to a named partner rather than guess. Clients booking a consultation rarely registered that the first exchange was handled by an AI employee. They registered that someone answered, immediately, and knew what they were talking about.

What a firm like this can expect

On the patterns we design for, a practice of this profile could expect to book around 31% more consultations within two quarters, and to trace roughly £148,000 of new instructions to enquiries that would otherwise have arrived cold on Monday. The partners would win back about eleven hours a week, because the qualifying happens before the call rather than on it.

The figure that tends to land hardest with partners is not the revenue. It is the response time the model turns on: a median of under a minute, day or night, rather than the working day a manual reply takes. For a firm willing to answer first, in commercial law, answering first is most of the work.

Could this work for your firm?

If enquiries reach you outside the hours a fee-earner is at a desk, the leak in this case study is almost certainly happening in yours. A thirty-minute call maps where your enquiries go cold and what an Autonomous Digital Branch would return. No pitch deck, just the numbers.