Modernising a Professional Services Firm Without Disrupting Billable Work

The hardest thing about modernising a law firm, an accountancy practice or a consultancy is not the technology. It is that everyone who would need to be involved is chargeable, and every hour they spend on an internal project is an hour nobody pays for.

This produces a specific pattern. The project is agreed enthusiastically in January, the people who matter are too busy to attend the workshops, and by September it has quietly stopped. Nobody killed it. It simply lost to the work.

This article is part of our guide to modernising a legacy business.

What a professional services firm actually sells

It sells judgement, and the record of that judgement is scattered across three places: a practice management system that records what was billed, an email archive that records what was actually decided, and the heads of the people who decided it.

The first is usually fine. The second is where the value is and it is effectively unsearchable. The third walks out at five o'clock and retires eventually.

That framing changes what to modernise. The instinct is to replace the practice management system, because it is the thing everyone complains about. It is rarely the highest-return move.

The four problems, in order of return

1. The same information is captured three times. A matter is opened in one system, the client details are typed into a document template, and the billing narrative is written from scratch at month end from someone's recollection. Each step is small. Multiplied across a firm it is the largest single cost.

2. Precedent is unfindable. Every firm has drafted the clause, built the model, or written the report before. Finding the previous version depends on somebody remembering it exists. Junior staff, who most need it, are least likely to know.

3. Nobody knows which work is profitable. Time is recorded against matters. Whether a matter type is systematically underquoted is a different question, and most firms cannot answer it. See your pricing history is a dataset.

4. Onboarding takes too long. A new fee earner takes months to become useful, mostly absorbing undocumented judgement. That period is a direct cost and it recurs with every hire.

Working around the billable hour problem

This is the actual constraint and it deserves designing around rather than complaining about.

Never ask for a workshop. Ask for forty minutes, individually, at a time they choose. Six forty-minute conversations produce more than one three-hour session that three of the six will miss.

Use the work as the data. You do not need people to describe their process if you can look at what they actually produced. Matter files, time entries and email threads describe the process more honestly than an interview would.

Deliver something in the first six weeks. Fee earners believe internal projects do not finish, because in their experience they do not. One visible improvement early buys the credibility for the rest.

Pick a sponsor who bills. A project owned entirely by operations will be politely ignored. It needs someone chargeable who wants it.

What to do first

Start with billing narratives or matter opening. Whichever consumes more time in your firm. Both are pure duplication, both are universally disliked, and both can be improved without changing how anyone does the actual work. This is the six-week win.

Then make precedent findable. Not a new document management system. A way of asking a question in normal language and getting back the three most relevant previous documents with the reasoning attached. This is the single highest-value thing for a firm with a long history, and it is now considerably cheaper to build than it was. See internal knowledge chatbots.

Then look at profitability by matter type. A data project rather than a systems project. It usually finds one service line running consistently over.

Then consider the practice management system, and frequently conclude it is fine and just needs connecting to things. Replacing it is the most disruptive option available and it is rarely the one that returns most. See rip and replace vs modernise around the edges.

The confidentiality constraint

Everything above runs into professional obligations before it runs into data protection, and that is worth handling early rather than discovering late.

Engagement letters commonly restrict use of client information to the purposes of that engagement. Professional rules impose duties beyond the statutory ones. Neither necessarily prevents a precedent system or a profitability analysis, and both determine how it must be built: access controls that respect matter boundaries, aggregation where insight crosses clients, and clear records of what was used for what.

Establish this in week one. It changes the design, and finding out in month four is expensive. See can you legally use your own customer data to train AI.

What it costs

For a firm of thirty to a hundred and fifty fee earners:

• Billing or matter-opening duplication removed: £6,000 to £18,000, payback usually inside a year
• Precedent search across an existing archive: £15,000 to £45,000, depending on the state of the documents
• Profitability analysis: £10,000 to £25,000
• Practice management replacement, if genuinely needed: £60,000 upwards, and considerable disruption

The first three together cost less than the fourth and return more.

Key Takeaways

• The constraint is the billable hour, not the technology. Design the project around it: individual forty-minute conversations, not workshops.
• What the firm sells is judgement, and the record of it is in the email archive rather than the practice management system.
• Start with billing narratives or matter opening. Pure duplication, universally disliked, improvable in six weeks.
• Making precedent findable is the highest-value move for a firm with history, and it has become much cheaper to build.
• Handle the confidentiality position in week one. It shapes the design and is expensive to discover in month four.

Frequently Asked Questions

Should we replace our practice management system?

Usually not as a first move, and often not at all. The common complaint is not that it records the wrong things, it is that nothing else talks to it and the interface is tired. Connecting it costs a fraction of replacing it and carries far less risk.

Can we use AI on client files?

Depending on your engagement terms, your professional rules and how the system is built, often yes. The determining factors are where the data goes, who can see what, and whether insight crossing clients is aggregated. A system built inside your own environment is a materially different proposition from pasting a matter into a public tool.

How do we get partners to engage?

Deliver something small that saves them time before asking for anything substantial. Partner engagement follows evidence, not plans. The firms where this works are the ones that spent the first six weeks on billing narratives rather than on a strategy document.

We tried this before and it failed. What was probably wrong?

In most cases, one of three things: it was scoped as a system replacement rather than a series of small improvements, it required workshops nobody attended, or it had no chargeable sponsor. All three are fixable and none of them are about the technology.

---

Running a professional services firm and unsure how to modernise without losing fee income? Talk to Halo Technology Lab. Our AI solutions for professional services start with the duplication that costs you most, and our work with Prezien took the same approach.