Every consulting firm owner I talk to says some version of the same thing. “We know we’re not billing everything we do. We just don’t know how much.” That gap between hours worked and hours billed is one of the biggest quiet drains on a firm’s profitability, and it doesn’t show up as a line item anywhere. It shows up as a margin that’s thinner than it should be, month after month.
For firms in the $1M to $25M range, we typically see $80,000 to $300,000 a year sitting in that gap. Not fraud, not laziness. Just the natural result of asking senior, busy people to remember and log every fifteen-minute email exchange, every client call that ran long, every “quick question” that turned into forty minutes of real advisory work.
Why timesheets don’t solve this
Most firms already have a timesheet system. That’s not the problem. The problem is what timesheets ask a human brain to do at the end of a long day: reconstruct, from memory, every billable moment across three, five, sometimes eight active clients.
Think about what a typical partner or senior consultant actually does in a given week. They answer client emails between meetings. They join a call that was scheduled for 30 minutes but ran to 55. They review a draft deliverable on a Sunday night and send back three pages of comments. They take a client’s call while walking to lunch and give real advice, off the clock, because that’s what relationship-driven work looks like.
None of that gets captured cleanly. By the time Friday afternoon timesheet entry rolls around, most people are reconstructing a rough approximation of their week. The easy, obvious blocks get logged. The scattered stuff, the fifteen-minute here-and-there work that actually adds up to hours, gets rounded down or forgotten entirely.
This isn’t a discipline problem you fix with a stricter policy. It’s a capture problem. You’re asking humans to do something computers are much better at, which is noticing and recording what actually happened.
What unbilled time actually costs a firm your size
Let’s put real numbers against this. Say you have 12 client-facing staff, billing an average rate of $200 an hour, and each person is under-capturing just 90 minutes a week of legitimately billable work. That’s not a dramatic number. It’s the kind of leakage that feels invisible in any given week.
Across 12 people, 90 minutes a week works out to roughly 940 hours a year of unbilled work. At $200 an hour, that’s close to $188,000 in revenue your firm did the work for and never invoiced. And that’s before you factor in the senior partners whose time is worth more and whose calendars are the most fragmented.
That number matters because it’s not a hiring problem or a pricing problem. It’s an operations problem, and it’s fixable without adding headcount or raising rates.
The manual work this use case actually targets
When we sit down with a consulting firm to map this out, the unbilled time tends to cluster in a few predictable places.
Email and Slack advisory work. A client emails a question. A partner answers with three paragraphs of real strategic guidance. Nobody logs it because it doesn’t feel like “work,” it feels like being responsive. Multiply that by every client, every week.
Meetings that run over or weren’t on the billing radar at all. Internal syncs that drift into client strategy. Calls booked as “quick check-in” that become 45 minutes of scope-relevant discussion. The calendar entry says one thing, the actual conversation did something else entirely.
Document review and revision. Someone spends two hours on a Sunday marking up a client deliverable. It happens outside work hours, outside the normal rhythm of the day, and it’s exactly the kind of work that gets skipped when Monday’s timesheet reminder shows up.
Cross-project bleed. A senior person answers a question for Client B while technically working on Client A’s engagement. Without a system tracking context in real time, that ten-minute detour just disappears.
None of this is anyone’s fault. It’s the natural outcome of knowledge work that happens in fragments, across channels, without a system built to notice those fragments as they happen.
What an AI agent doing this actually looks like
Here’s where the shift happens. Instead of asking your team to remember and log their time, you build a system that watches the actual work product and reconstructs the time picture for them.
An agent built for this reads the sources where billable work already leaves a trace: email threads tagged to a client, calendar entries and their actual duration versus scheduled duration, meeting transcripts from calls (most firms are already recording these for notes), and document activity like edits, comments, and reviews tied to a client folder or project code.
It then does the matching work a human would do manually, except it does it continuously and without fatigue. It maps activity to the right client and matter code. It flags a 30-minute scheduled call that ran to 58 minutes as a discrepancy worth reviewing. It picks up the Sunday night deliverable review and drafts a time entry for it before Monday morning. It notices that an email thread with Client B contains substantive advisory content, not just scheduling logistics, and drafts an entry accordingly.
Critically, it doesn’t auto-bill anything. It drafts. A person still reviews and approves every entry before it hits an invoice. What changes is the starting point. Instead of a blank timesheet and a tired memory, your team is reviewing a pre-populated draft that’s already 80-90% accurate, adjusting the last details, and moving on. The reconstruction work, the hardest and most error-prone part, is already done.
This is the same underlying approach we use in Omni’s other consulting agents, just pointed at a different problem. The Knowledge Agent we build for firms reads every deck, document, and meeting transcript a firm produces so people can ask questions across the whole corpus instead of hunting through old project folders. A time-capture agent uses that same reading capability, applied to the question of what actually happened this week and who should be billed for it.
The Research Agent we build for the start of every new engagement works on a similar principle. It replaces weeks of repeated secondary research with a structured brief, sources included. The Proposal Generation Agent does the same for pitch work, pulling from past proposals and case studies instead of starting from a blank page every time. All three exist because the same pattern shows up again and again in consulting firms: valuable work is happening, but it isn’t being captured, reused, or billed the way it should be.
Why this compounds beyond billing accuracy
Fixing time capture doesn’t just recover revenue. It changes how your firm understands its own profitability.
Right now, if you ask “which clients are actually profitable once we account for all the time we put in,” most firm owners can’t answer that with confidence. The data is too thin, because the underlying time entries were never accurate in the first place. Once you’re capturing time closer to what actually happened, you get a real read on realization rates, on which engagement types quietly eat more senior hours than they bill for, and on where your pricing is misaligned with actual effort.
That’s the kind of insight that should inform your next round of proposals and rate cards, not just your next invoice run. It’s also exactly the kind of operational clarity we talk through during the AI audit for consulting firms, because billing accuracy and firm-wide knowledge management debt tend to be two sides of the same coin. The firm that can’t find its own past work product is usually the same firm that can’t accurately account for the time spent producing it.
If you want a structured way to think through where your first agent should go, we put together a practical worksheet called Deploy Your First Business Agent. It walks through how to pick the right first use case, what data sources you already have, and what “done” looks like for a pilot. It’s a useful next step whether time capture is your top priority or something further down the list like proposal work or research.
What the Omni Audit actually delivers
We built the Omni Audit specifically because most firm owners don’t have 40 hours to spend evaluating AI vendors, sitting through demos, and building a business case from scratch. The audit is 60 minutes, on a call, and it produces three concrete things.
First, a map of where your billable and unbilled time is actually going right now, based on how your team currently works, not a generic industry assumption. Second, a rough dollar estimate of what’s leaking, using your real headcount and rates rather than a benchmark that may not fit your firm. Third, a short list of the one or two agents that would move the needle fastest for your specific situation, whether that’s time capture, proposal generation, research, or knowledge management.
No slide deck, no sales pitch dressed up as a workshop. Just a clear picture of the problem and a realistic next step. You can book my Omni Audit directly if you want to see what that looks like for your firm.
Getting started without disrupting how your team works
One thing worth saying plainly: you don’t need to replace your existing time tracking or billing software to do this. The agent sits alongside what you already use. It reads from your email, calendar, and document systems, and it writes drafts into your existing timesheet or practice management tool. Your team’s workflow barely changes on the surface. What changes is how much of the manual reconstruction they still have to do themselves, which for most firms drops from “almost all of it” to “just the final check.”
We usually recommend starting with one team or one practice area rather than rolling it out firm-wide on day one. That gives you a real read on accuracy and adoption before you commit further, and it means the first win is visible and specific rather than diffuse. A trades-business owner in our broader network describes this same pattern from a different angle. Once you can see the gap between logged hours and actual hours, on a single team, the case for expanding the fix becomes obvious pretty fast.
If you’re weighing this against other operational fixes, it’s worth reading more broadly about how firms your size are sequencing AI adoption. Our guides section has more detail on how the Research Agent and Knowledge Agent work in practice, and our blog covers how other consulting and advisory firms have approached their first agent rollout. For a broader view of where this fits inside a full advisory operating model, Omni Advisory is worth a look too.
The real decision in front of you
The math here isn’t complicated. If your firm is losing even a conservative estimate of $80,000 to $150,000 a year to time that never gets captured, the cost of not fixing this is far higher than the cost of a focused pilot. And the fix doesn’t require a system overhaul or a change in how your consultants actually work day to day. It requires pointing the right tool at the problem.
The best way to know your specific number, rather than an industry range, is to walk through it directly. See Omni for consulting firms and get a clear, numbers-based picture of what unbilled time is actually costing your firm this year. Or skip straight to the conversation and book a 60-min Omni Audit to see exactly where your firm stands.