You bill 1,800 hours a year. Your associate bills 1,600. But the actual work you both do is closer to 2,100 hours each. The gap isn’t laziness. It’s the five-minute call you took at 4:30 PM, the 20-minute email thread you answered from home, the document review you squeezed in between depositions. None of it made it into your time-entry system because you were already thinking about the next matter.

For a two-partner firm with three associates, that leakage typically runs $80,000 to $250,000 annually. Most of it happens in the same places: intake conversations that never get logged, client emails answered outside billing windows, document prep that feels too small to record. The pattern is predictable. The fix isn’t more discipline. It’s removing the need for manual entry altogether.

Why attorneys don’t log time

The standard advice is to enter time as you go. Block 15 minutes at the end of each day. Use a timer. The problem is that advice assumes your day has natural breaks. It doesn’t. A client calls during lunch. You draft a motion outline in the parking lot before a hearing. You review a contract at 9 PM because the closing is tomorrow. None of those moments include a pause to open your billing software and describe what you just did.

The bigger issue is context-switching cost. Every time you stop to log an entry, you’re pulling yourself out of the substantive work. For a six-minute task, the logging takes two minutes. You’ve just added 33% overhead to a billable activity. Do that 40 times a week and you’ve spent an hour and a half on administrative work that generates zero revenue. The math doesn’t close.

Junior associates face a different version of the same problem. They’re not sure what’s billable. A partner asks a question in the hallway. Is that .1 hours? The client sends a one-line email. Do you bill for reading it? The uncertainty leads to under-reporting. One associate in our network described losing an entire morning’s work because she couldn’t reconstruct what she’d done across four different matters by the time she sat down to enter time at 3 PM.

What real-time capture looks like

AI agents solve this by watching the work as it happens. They don’t wait for you to remember. They parse your email, listen to your calls, and scan the documents you open. When you spend 12 minutes drafting a response to opposing counsel, the agent logs it. When you review a purchase agreement for 40 minutes, the agent captures the file name, the time span, and the matter code. When you take a call from a client at 7 PM, the agent records the duration and attaches it to the right billing entry.

This isn’t theoretical. We build these systems for firms that bill $150 to $600 per hour. The Matter Triage Agent sits on your inbox and classifies every inbound message by practice area and urgency. It knows which emails are client communication (billable) and which are internal admin (not billable). It drafts the time entry in your billing system’s format and queues it for one-click approval. You review a list at the end of the day, approve 12 entries in 90 seconds, and you’re done.

The Intake Voice Agent does the same thing for phone calls. Every inbound call gets answered, even at 6 PM on a Friday. The agent asks the caller about their matter, runs a conflict check, and books a consultation. It also logs the call duration and attaches a transcript to the matter file. If the call qualifies as a billable intake consultation under your fee structure, the entry is already written. If it’s a prospect who didn’t retain, the agent marks it non-billable and files the transcript for follow-up. You can read more about how this works at the AI audit for law firms.

Document work is where the time leakage gets expensive. A senior associate spends three hours reviewing a commercial lease. She makes notes in the margin, highlights clauses, and drafts a two-page memo. But she forgets to log the first 90 minutes because she started the review during a gap between meetings and didn’t think to start a timer. The Document Review Agent tracks which files you open, how long they’re active, and whether you made edits. It writes a draft entry: “Review and analysis of commercial lease agreement, [file name], 3.2 hours.” You approve it. The client gets billed. The revenue doesn’t disappear.

The three places firms lose the most billable time

Client communication outside office hours

A client emails you at 8 PM with a question about a settlement offer. You read it on your phone, think through the answer, and reply with three paragraphs of analysis. That’s 15 minutes of billable work. But you’re at home. You’re not logging into your time-entry system. By Monday morning, you’ve forgotten the exchange happened.

An agent monitoring your email sees the thread, measures the time between when you opened the message and when you sent the reply, and logs the entry automatically. It pulls the subject line, the client name, and the matter number from your practice-management system. The entry appears in your queue the next morning. You confirm it’s accurate and move on. The client gets billed for the advice you gave. You didn’t lose the revenue to memory decay.

Intake and conflict checks

Every intake call has billable components. The initial consult. The conflict check. The follow-up email confirming next steps. But if the caller doesn’t retain, many firms write off the time entirely. That’s a choice, but it should be a deliberate one. More often, the time just never gets recorded because the intake felt like business development rather than client service.

The Intake Voice Agent treats every call as a potential billable event. It logs the call, records the conflict-check result, and drafts an entry. If your firm’s policy is to bill initial consults, the entry is marked billable. If your policy is to write them off unless the client retains, the agent marks it non-billable but keeps the record. Either way, you have a complete audit trail. You’re not guessing what happened when you review your pipeline three weeks later.

We’ve also built a practical worksheet that walks through the intake-automation decision tree. If you want to map your current process against what an AI-native intake flow looks like, download the AI Client Intake Checklist for Law Firms. It’s a one-page tool you can use in a partner meeting to identify where your intake leakage is happening.

Document prep and research

An associate spends two hours researching a motion. She reads four cases, pulls relevant quotes, and drafts a memo. She logs 1.5 hours because she’s not sure if the time spent reading background material counts as billable research or as her own learning. The uncertainty leads to under-reporting.

The Document Review Agent doesn’t guess. It tracks which case files were open, how long each was active, and whether the associate copied text into a work product. It drafts an entry: “Legal research re: motion to dismiss, review of [case names], 2.0 hours.” The associate reviews it, confirms it’s accurate, and approves. The client gets billed for the full scope of work. The associate doesn’t have to make a judgment call about what qualifies.

This is especially valuable for discovery review. A junior associate spends six hours reviewing 300 pages of contracts. She flags 40 clauses, writes summaries, and creates a privilege log. But she only logs four hours because she’s not confident the client will pay for the full time. The partner never sees the discrepancy. The firm eats two hours of billable work. The agent logs all six hours, attaches the output, and lets the partner decide whether to adjust the entry. The default is full capture. Write-offs become visible decisions, not invisible leakage.

What an Omni Audit shows you

We run a 60-minute diagnostic with three deliverables. First, a time-capture map that shows where your billable hours are disappearing. We pull anonymised data from your email, calendar, and document activity over a two-week sample. We don’t need access to client files. We’re measuring time patterns, not reading privileged content. The output is a breakdown: how much time your team spends on email, calls, document work, and admin. How much of that time is currently captured in your billing system. And how much is leaking.

Second, an agent design for your highest-value use case. If intake is your biggest problem, we spec the Intake Voice Agent and show you what the call flow looks like. If document review is the bottleneck, we design the Document Review Agent and show you the first-pass output it would produce. If email triage is eating your day, we spec the Matter Triage Agent and show you how it routes messages and drafts entries. You see the agent working on real examples from your practice.

Third, a build-and-run cost model. We estimate the one-time build cost, the monthly run cost, and the annual value capture. For a five-attorney firm losing $120,000 a year to unbilled hours, the payback period is typically 90 to 120 days. You’re not buying software. You’re buying an operational system that runs inside your existing stack and pays for itself in recovered revenue.

Book a 60-min Omni Audit and we’ll walk you through the numbers for your firm. You’ll leave with a capture map, an agent spec, and a cost model. No deck. No sales pitch. Just the three outputs you need to make a decision.

How this connects to your billing system

The agents don’t replace your practice-management software. They feed it. If you’re running Clio, the agent writes entries in Clio’s format and pushes them to your draft queue. If you’re using PracticePanther or MyCase, the agent adapts to that schema. You review and approve entries in the same interface you use today. The only difference is that the entries are already written when you open the system.

This matters because most firms have years of workflow built around their billing platform. You’re not going to rip that out. You’re not going to train your team on a new system. The agent is a data layer that sits upstream. It watches your work, drafts the entries, and submits them to your existing approval process. Your workflow doesn’t change. Your capture rate does.

One litigation partner described the shift this way: “I used to spend 20 minutes at the end of every day trying to remember what I did. Now I spend three minutes approving entries that are already 90% accurate. I’m billing an extra four hours a week just from better capture. That’s $1,200 a week I was leaving on the table.”

Why firms wait and what it costs them

The most common objection is that the team will adapt the system to inflate hours. That’s a governance problem, not a technology problem. If an associate is logging time they didn’t work, you’ll see it in the output quality. If a partner is approving inflated entries, you’ll see it in client complaints and write-offs. The agent doesn’t create ethical risk. It makes the time record more complete and more auditable.

The second objection is client pushback. Some clients will question a 3.2-hour entry for document review when they’re used to seeing round numbers. The answer is that 3.2 hours is what the work took. If the client wants a discount, that’s a pricing conversation. But the starting point should be an accurate record of the work performed, not a pre-discounted guess. The agent gives you the accurate record. You still control the final bill.

The cost of waiting is straightforward. If your firm is losing $150,000 a year to unbilled hours, every quarter you delay costs you $37,500 in revenue that’s already been earned but never invoiced. That’s not opportunity cost. That’s work your team did, value your clients received, and money you’ll never collect because the time entry was never written.

You can explore more about how Omni handles operational automation at Omni Ops, or see how voice agents handle intake at Omni Voice. If you want a broader view of how AI is changing professional services, the EDNA insights library covers the operational patterns we see across verticals.

What happens after the audit

If the numbers work, we build the agent in 10 to 15 business days. You get a staging environment where you can test the agent on historical data before it touches live client work. We train it on your matter codes, your billing policies, and your approval workflow. We connect it to your email, your phone system, and your document storage. We don’t ask you to change how you work. We ask you to review draft entries and approve the ones that are accurate.

The agent goes live in observation mode first. It logs entries but doesn’t submit them. You review a week’s worth of output and tell us what needs adjustment. We tune the classification rules, the time-rounding logic, and the matter-matching algorithm. Once you’re confident the entries are accurate, we flip it to active mode. The agent starts submitting entries to your billing system. You approve them. Your clients get billed. Your realization rate climbs.

Most firms see a 15 to 25 percentage-point improvement in time capture within the first 60 days. That’s the difference between billing 70% of the work you do and billing 90%. For a firm doing $2 million in annual billings, that’s $400,000 in recovered revenue. For a firm doing $8 million, it’s $1.6 million. The agent doesn’t make you work more hours. It makes sure the hours you’re already working show up on the invoice.

See Omni for law firms to understand what the full diagnostic process looks like, or book my Omni Audit and we’ll run the numbers for your practice. You’ll know within an hour whether this is worth building. If it is, we’ll have the first agent running in two weeks. If it’s not, you’ll have a capture map that shows you exactly where your time is going and what it would take to fix it manually.

The work is already happening. The question is whether you’re getting paid for it.