Small law firms

You bill about three hours a day. Running the firm eats the other five.

Firms from solo practice to about fifty people, in any practice area: family, immigration, employment, personal injury, criminal defense, trusts and estates. The loops on this page run the same way in every one of them. Where a practice deserves its own deep dive, it gets its own page, and probate and estate planning already has one.

The loop that bleeds

A firm of one to fifty people sells expertise by the hour or by the flat fee, and the average lawyer captures 2.9 billable hours in an 8-hour day. The rest goes to intake, conflicts checks, status updates, time entry, and billing itself. There is no operations department to hand any of it to, so the highest-rate person in the building is the operations department.

The work that would fill the billable day leaks at the front door. When Clio ghost-shopped 500 firms as a prospective client, a third replied to the email and 40 percent answered the phone. A prospect who hears nothing for two days hires the firm that answered in ten minutes, and the matter that would have paid for the month walks with them.

The obvious fix is the one thing a law firm cannot do casually. Client files are confidential, ABA Formal Opinion 512 keeps the competence and supervision duties on the lawyer whatever tool is in the loop, and most consumer AI products give no usable account of where the data goes. So the smallest firms, the ones carrying the heaviest administrative load per lawyer, adopt at the lowest rate, and the five unbilled hours stay.

2.9 hours: Billable time the average lawyer captures in an 8-hour day, a utilization rate of 37 percent. The other five hours are firm overhead, worked by its most expensive people.
33%: Share of 500 ghost-shopped firms that replied to a prospective client’s email. 40 percent answered the phone. Both numbers fell between 2019 and 2024.
24%: AI adoption at firms of two to nine attorneys, against 30 percent across all firm sizes. The confidentiality constraint lands hardest on the firms with the least slack.

Where to start

Three places this pays for itself first.

Each one is fixed scope, ships in weeks, and either recovers money already earned or removes the thing blocking the work after it. Start wherever it hurts most. None of them requires you to sell more or change what you do.

Billable-hour leak capture

What breaks

Time entry happens at the end of the week or the end of the month, reconstructed from the calendar and the sent folder. Reconstruction has a known direction: the two-hour drafting block survives, and the six-minute email does not, even when there were eight of them that day. The shortfall repeats per lawyer, per day, and it never shows up on a report, because every report is built from the entries that did get made.

What gets built

Read the exhaust the firm already produces: calendar, sent mail, call log, and document edit history. Draft time entries per matter, narratives included, and put them in front of each timekeeper the next morning as a proposed day to accept, edit, or strike. Nothing reaches a prebill without that review. The deliverable is a complete first draft of the day, built from records instead of memory.

What already solves this

Clio, MyCase, and PracticePanther all ship timers, and a timer captures time when someone remembers to start it, which is the exact failure being fixed. Laurel, formerly Time by Ping, sells passive capture to large firms, priced and integrated for a document management stack a ten-person firm does not run. TimeSolv and Toggl are entry tools with better ergonomics and the same dependence on the person typing. Meanwhile the raw material already sits in the firm’s own Microsoft 365 or Google tenant, which is what makes this buildable at small-firm cost.

Opens onto: Prebill review, collections follow-up, and the utilization report the partners have never had.

The day gets drafted from records. The lawyer reviews it instead of reconstructing it.

5 hours

The nonbillable remainder of the average lawyer’s day. Some of it is management that will never bill. Enough of it is six-minute increments that were worked and never entered.

Derived from Clio 2024 Legal Trends Report utilization data

Intake and conflicts pipeline

What breaks

Inquiries land in a personal inbox, a voicemail box, and a website form nobody owns, and the person expected to reply is whoever is deepest in a filing deadline. The conflicts check runs on a name search and a partner’s memory, so even an inquiry the firm wants waits days for an engagement letter. The inquiries the firm never answers appear nowhere as a loss, which is why the front door stays broken.

What gets built

One queue for every inquiry, whatever door it came through. Each entry gets a conflicts sweep against the full matter history, the contact list, and the email archive, a drafted reply in the firm’s voice, and a drafted engagement letter when the sweep comes back clean. An attorney reviews everything before it leaves the building. The queue also produces the number nobody has today: how many inquiries arrived this month, and how many got an answer.

What already solves this

Clio Grow, Lawmatics, and Smokeball run intake pipelines well, and if one is set up and worked at your firm, keep it. What they check for conflicts is the data somebody previously typed into them, and a small firm’s history lives in closed matters, old email, and the memory of whoever was there in 2019, none of which an intake CRM has read. Ruby and Smith.ai answer the phone with a person, which fixes the pickup rate and cannot run a conflicts sweep or quote your process. The knife: every one of these tools assumes the firm’s history is already structured, and at most firms it is not.

Opens onto: Engagement letters and fee agreements as a pipeline, matter opening, and the status updates that generate most inbound client calls.

One queue, a conflicts sweep, an answer. The dotted box stops filling.

2%

Share of ghost-shopped firms whose email reply mentioned experience with similar cases, the one thing the shopper had asked about.

Clio 2024 Legal Trends Report

Locally hosted drafting and review

What breaks

Every practice has a recurring document set: petitions, discovery responses, demand letters, declarations, the same exhibits assembled the same way. Drafting them is the obvious place for AI assistance, and the confidentiality duty makes consumer tools a nonstarter, because client files cannot go into products that give no usable account of retention or training use. So the drafting stays manual and the firm reads about AI instead of using it.

What gets built

Drafting and review assistance on infrastructure the firm controls: a local model on firm hardware, or a private tenant with a written data-handling posture your engagement letters can survive. It reads the matter file where it already lives, produces first drafts of the recurring set in the firm’s own prior language, and routes every draft to the responsible attorney. Formal Opinion 512’s competence and supervision duties are design constraints here, the same way they are in any delegation.

What already solves this

CoCounsel and Harvey are the reference products, built and priced for firms far larger than fifty people, and both are cloud services. Spellbook lives in Word and drafts transactional documents well, also in the cloud. Whether any of them fits depends on what your engagement letters and protective orders say about where client material may go, and that check comes before any tool choice. What nobody sells is the version that runs where the client file lives and never sends it anywhere, because there is no subscription revenue in it.

Opens onto: The practice-specific pipelines: county filing sets, discovery production, immigration form packages, whatever document set your matter type repeats.

The model comes to the file. The file never crosses the firm boundary.

Op. 512

The ABA’s generative AI opinion. Confidentiality, competence, and supervision stay with the lawyer using the tool, whatever the vendor’s marketing says.

ABA Formal Opinion 512, July 2024

What I would ask you

I know the shape of this. I don't know your version of it.

These are the questions that decide whether any of the above is worth building at your firm, and they get asked before anything is scoped. Several of them can kill the project, which is the point.

What happens to an inquiry that arrives Thursday at 4pm, and who can say by Friday morning that it exists?
How does a conflicts check run today, and does anything search the closed matters and the email archive, or only the contacts somebody once typed in?
Pick last month. How many hours did the firm work, how many were entered, and how many were billed?
Who writes the time entries, when, and how much of that is reconstruction from the calendar?
What do your engagement letters promise clients about their files, and does any tool the firm uses today already break that promise?
Which document does the firm produce most often, and how much of it changes from one matter to the next?

Where it goes

A first win buys the right to the bigger one.

Start here

The billable-hour leak capture. It reads records the firm already produces, needs write access to nothing, and the recovered increments in the first prebill cycle are the business case for everything after it.

Then the front door

The intake and conflicts pipeline, because the fastest way to fill the billable day the first build measured is to stop losing the matters that ring the firm on their own.

Then the platform

Locally hosted drafting, once the firm has watched reviewed output for a quarter and knows what its supervision workflow looks like. By then the recurring document set has picked itself.

Every figure on this page comes from industry research. None of it comes from my own engagements. I have not put a client's results here, because I am not going to dress up someone else's benchmark as my track record. When there is a delivered number worth showing, it will appear here with the client's name on it or not at all.