§ Case Studies

Three AI Agents, One Solicitors' Practice, 47% More Client Matters Handled

Luke Needham··9 min read
Three AI Agents, One Solicitors' Practice, 47% More Client Matters Handled

Barlowe & Associates is a six-person employment law practice in Leeds. At the start of 2026, the three solicitors were billing around 70% of their available hours — not because work was scarce, but because 27 hours a week were disappearing into drafting, client intake, and matter management admin. Three AI agents changed that picture entirely. Sixty days later, the practice was handling 47% more client matters with the same headcount, 29 hours of weekly admin had been automated, and the solicitors were billing at 91% capacity. Here is exactly what we built.

The Constraint: 27 Hours a Week Lost to Admin

A busy UK solicitors office with stacks of legal documents and case files representing the 27 hours of weekly admin consuming billable time

When we ran the initial time audit at Barlowe & Associates, the numbers were stark. Across the three solicitors and the practice manager, 27 hours a week were going into work that was not legal work. Specifically:

  • 9 hours drafting standard documents — employment contracts, settlement agreements, tribunal response letters, and without prejudice letters that followed the same structural pattern every time
  • 7 hours on client intake — responding to initial enquiries, qualifying cases, gathering background information, and scheduling consultations
  • 6 hours on matter management — sending status updates, chasing outstanding client documents, logging time manually, and preparing billing summaries
  • 5 hours on internal administration — meeting notes, file management, and general communications

The partner leading the practice, Daniel Barlowe, described the situation plainly: "We turn away or delay good cases because we are too bogged down in the paperwork on the ones we already have. We are not short of legal work. We are short of time to do it."

This is the capacity constraint that shows up in almost every professional service firm we audit. The work is there. The expertise is there. What is missing is the operational bandwidth to handle more of it. As we found in our work with a four-person Edinburgh IFA practice, the ceiling is almost always administrative rather than professional. The solution is the same: identify the three to five highest-volume admin processes and build an agent for each one.

The problem was not the quality of the legal work. The problem was that every hour of legal work came bundled with roughly 40 minutes of administrative work that a solicitor was overqualified to do.

Agent One: The Document Drafting Agent

A solicitor reviewing an AI-generated legal document draft on screen, showing the drafting agent workflow for employment contracts and settlement agreements

The highest-value agent was the one that addressed the biggest time drain: document drafting. Employment law practices produce the same set of documents repeatedly — contracts of employment, settlement agreements, without prejudice letters, tribunal response bundles, disciplinary outcome letters, and redundancy notices. The structure of each document type is fixed. What varies is the specific parties, facts, and applicable clauses.

The document drafting agent works from a structured knowledge base containing: clean versions of all standard document templates, approved clause libraries for common scenarios (gross misconduct, performance management, TUPE, discrimination claims), precedents from previous successful matters, and the practice's preferred drafting style and language.

When a solicitor opens a new matter, they complete a brief matter information form — client name, employer name, matter type, key facts, relevant clauses to include. The agent reads this form and produces a first draft within three to four minutes. For a standard settlement agreement that previously took 90 minutes to draft from scratch, the agent produces a draft that requires 15–20 minutes of solicitor review and customisation.

We built the knowledge base using the practice's own documents — 60 precedents drawn from matters closed in the previous two years, cleaned and structured as a vector store. As we covered in the piece on AI agent memory architecture, the quality of the retrieval layer determines the quality of the output. The agent does not hallucinate clauses because it draws from a bounded library of pre-approved precedents, not from the open internet.

The SRA's requirements around AI use — that firms retain professional oversight and cannot rely entirely on AI output — are naturally satisfied by the workflow: every draft gets a solicitor review pass before it leaves the practice. The agent removes the repetitive production work; the solicitor retains the professional judgement about what the document should say.

Time saving: 9 hours per week reduced to under 2 hours. 7 hours recovered.

Agent Two: The Client Intake Agent

A professional team using an AI-powered client intake system, showing how prospective employment law clients are qualified and onboarded automatically

Client intake at most solicitors' practices is a slow-motion bottleneck. A prospective client contacts the firm. Someone eventually responds — usually hours later, sometimes a day later. A back-and-forth begins to establish whether the case is within the firm's remit, whether the client can afford private fees or qualifies for legal aid, and what the relevant facts are. By the time this is complete, three to five days have passed and the prospect may have gone elsewhere.

The intake agent runs 24 hours a day. When a prospective client fills in the contact form on the practice website, the agent immediately sends an acknowledgement and begins a structured qualification conversation by email. The agent asks about the nature of the employment matter, the timeline of key events, whether the client is currently employed or has already left, whether ACAS early conciliation has begun, and the type of outcome the client is seeking.

Based on the responses, the agent categorises the matter — unfair dismissal, discrimination claim, settlement negotiation, disciplinary advice, TUPE, or other — and assesses whether it falls within the practice's scope. If it does, the agent books an initial consultation through the practice's scheduling system and creates a pre-populated matter file with the information gathered. If it does not, the agent explains why and, where possible, signposts appropriate alternatives.

By the time the solicitor sits down for the initial consultation, they have a two-page brief about the client's situation prepared by the agent. What was a 45-minute exploratory call becomes a focused 25-minute consultation. The solicitor arrives informed, the client feels heard, and the practice can run more consultations per day without any additional time investment.

This mirrors the client onboarding approach we described in the client onboarding tutorial: the agent's job is to do the information gathering so that the professional can spend their time on the professional part of the relationship.

Time saving: 7 hours per week of intake handling reduced to under 1 hour. 6 hours recovered.

Agent Three: The Matter Management Agent

The third agent handles the ongoing operational layer of live matters: status updates to clients, document chase-ups, time logging, and billing preparation. This is the agent that most practices underestimate — but it is the one that has the largest impact on client satisfaction and cash flow.

The matter management agent connects to the practice management system via MCP integration, reading the current status of each open matter and identifying which clients are due a status update, which outstanding documents have been waiting more than five days, and which matters have billable time that has not yet been logged.

Every Monday morning, the agent sends each active client a brief, plain-English update on their matter — what happened last week, what is happening this week, and what the client needs to do (if anything). This alone eliminated approximately 80% of inbound client enquiries asking "what is happening with my case?" — the most common and most time-consuming category of client communication in a busy practice.

For time logging, the agent reads completed tasks and drafted documents and creates provisional billing entries that the fee-earner reviews and approves. The practice had previously estimated that around 15% of billable time went unlogged due to the friction of manual time capture. The agent now catches it consistently.

Time saving: 11 hours per week of matter management admin reduced to under 2 hours. 9 hours recovered.

The Results After 60 Days

A business analytics dashboard showing the 60-day results of the AI operating system deployment at Barlowe and Associates — 47% more matters, 29 hours recovered, billing capacity at 91%

Sixty days after deployment, the numbers at Barlowe & Associates looked like this:

  • 47% more client matters handled per month — from an average of 34 active matters to 50, same six-person team
  • 29 hours of weekly admin automated across the three agents — drafting (7h), intake (6h), matter management (9h), residual admin (7h)
  • Billable capacity up from 70% to 91% — the 27-hour admin burden reduced to under 5 hours, freeing 22 hours a week for client work
  • Consultation-to-instruction conversion up 18% — faster response time and better-prepared initial consultations improved the rate at which prospects became paying clients
  • Client satisfaction score up from 7.1 to 8.8 out of 10 — weekly matter updates eliminated the "why hasn't anyone contacted me?" complaint that dominated the practice's feedback
  • Estimated 15% more billable time captured through automated time logging, increasing revenue without any additional work being done

Daniel Barlowe's assessment after 60 days: "We have effectively added a full-time operational layer to the practice without hiring anyone. The agents handle the work around the legal work — the intake, the drafting, the updates, the billing admin — so we can focus on the actual legal work."

The agents do not practise law. They do everything around the law that used to prevent lawyers from practising it.

What This Cost and What It Pays Back

A professional UK legal services team, representing the outcome of deploying an AI operating system — more capacity, better client service, and a stronger business

The total build cost for the three-agent AI operating system at Barlowe & Associates was £3,800 — a single engagement covering the knowledge base setup, agent configuration, practice management system integration, testing, and a 30-day post-deployment refinement period.

Ongoing running costs are approximately £85 per month: cloud compute, language model API, vector store, and scheduling integration. There is no per-seat licensing. The system runs continuously at a fixed operational cost.

The return on that investment came from three directions:

  1. Increased capacity. 47% more matters handled per month, at the same fee rates, represents a significant revenue increase from the same headcount. For a practice billing at an average of £1,800 per matter, the additional 16 matters per month represents £28,800 in additional monthly revenue potential.
  2. Recaptured billable time. The 22 hours per week converted from admin to billable time, at £180 average hourly rate across the team, represents approximately £15,840 in additional billable capacity per month — trending toward full utilisation as the practice grows into its new capacity.
  3. Reduced write-offs. The estimated 15% improvement in time capture has reduced monthly write-offs materially — the practice's single largest source of invisible revenue leakage.

The £3,800 build cost was recovered within the first week of full operation. The pattern is consistent with what we have seen in recruitment agencies, accountancy firms, and IFA practices: the constraint in professional service firms is almost always operational capacity, not market demand. Remove the operational constraint, and revenue follows.

If you run a solicitors' practice, a legal consultancy, or any professional service firm where qualified people are spending too much time on the work around the work — get in touch. We scope and build AI operating systems for UK service businesses in a single engagement, and we can tell you within a short call whether the numbers are likely to make sense for your practice.

L

Written by Luke Needham

Founder at Quantum Flow Automation — building AI systems that work.

§ 99Subscribe

More field notes, in your inbox.

One email per week. What we shipped, what broke, what's worth paying attention to in AI.

BOOK CALL