Every consulting firm has a version of this story. A partner is on a plane at 10pm, marking up a deck for a 9am client meeting. She’s not checking strategy at that point. She’s checking whether slide 14 uses last quarter’s numbers, whether the formatting matches the firm template, whether the case study on page 22 is even the right industry for this client.

That’s deliverable QA. It’s unglamorous, it’s invisible to clients when it’s done right, and it eats a shocking number of senior hours in firms doing $1M to $25M in revenue. Most owners never put a number on it because it’s spread across dozens of small reviews rather than one big line item. That’s exactly why it’s worth putting a number on.

The review nobody bills for

In a firm this size, the person doing final QA on a proposal or a deliverable is usually the most expensive person in the building. A partner, a principal, sometimes the founder. That’s who catches the outdated pricing table, the client name misspelled from a find-and-replace error, the chart that still says “DRAFT” in the footer.

None of that gets billed. It can’t. So it either happens on nights and weekends, or it gets rushed, or it slips through and the client notices before you do. We’ve talked to firms where a single bad deliverable, one with a wrong number in it, cost them a renewal conversation six months later. The QA problem isn’t really about typos. It’s about trust, and trust is the entire product in advisory work.

That range isn’t abstract. It’s built from three specific patterns we see over and over in firms your size.

Where the hours actually go

Proposal and pitch time. A senior team writing a major proposal from a blank page, pulling old decks from memory, guessing at pricing that worked last time. Twenty to forty hours is normal for a serious proposal, and most of that time is spent reconstructing things the firm already knows. The win rate might be fine. The cost of getting to “yes” is what’s brutal, and it’s the QA pass at the end, the partner checking the whole thing makes sense as one document, that eats the last few hours nobody planned for.

Research and synthesis. Every new engagement starts with two or three weeks of secondary research. Industry structure, competitor moves, regulatory context. A lot of it has been done before, for a different client in a similar sector, and nobody remembers it exists. So it gets redone, then it gets QA’d for accuracy against sources that may or may not still be current, and that review step is where a lot of quiet hours disappear.

Knowledge management debt. Every project produces real intellectual property. Frameworks, findings, slide language that landed well with a client. Almost none of it makes it into a form anyone else in the firm can find and reuse. So the firm pays for the same thinking twice, and when someone finally does try to reuse an old deck, a chunk of QA time goes into checking whether it’s still accurate, still on-brand, still true.

These three pains compound. A proposal built on stale research needs more QA. A deck built from unreused IP needs more fact-checking because nobody’s sure what’s current. The manual QA layer isn’t the root problem. It’s the symptom of a firm that has no system for keeping its own work current and reusable.

What this looks like with an agent doing the work

Here’s where it gets concrete. We build three agents inside Omni ops that directly attack this, and two of them work together in a way that changes how QA happens entirely.

The Proposal Generation Agent pulls from every past proposal, case study, and pricing structure the firm has ever produced, and drafts a tailored first version for the new opportunity. Not a generic template. A draft built from what actually worked, with the right case studies already matched to the client’s industry and the pricing already benchmarked against similar past deals. The partner’s job shifts from writing to reviewing, and reviewing a draft that’s already internally consistent is a fundamentally different task than reviewing something assembled from scratch under deadline pressure.

The Research Agent runs structured industry and company research the moment an engagement kicks off. It produces sources, summaries, and a one-page brief, and it does this the same way every time, which means the QA on it is fast because the format never changes and the sourcing is always attached. No more wondering if a stat in the deck came from a 2019 report someone forgot to update.

The Knowledge Agent is the one that closes the loop. It reads every deck, document, and meeting transcript the firm produces and can answer questions across the entire corpus. When someone’s building a new proposal and wants to know “have we done work like this before,” they get an actual answer, with the source deck attached, instead of a Slack message asking around the office. This is also your QA safety net. Before anything goes out the door, the Knowledge Agent can flag when a figure, a client reference, or a framework contradicts something the firm has said elsewhere.

Put together, these three agents don’t eliminate the partner review. They change what the partner is reviewing. Instead of checking spelling and pulling old files, the review becomes a genuine second look at strategy and fit, the part of the job that actually needs a partner’s judgment.

Why this matters more at $1M to $25M

Firms below this size don’t have enough deliverable volume for the problem to compound. Firms well above it usually have a dedicated knowledge management function, even if it’s underused. Your firm is at the size where the pain is real but a full internal solution isn’t. You’ve got enough proposals, enough client engagements, enough decks in enough shared drives that nobody has a full picture of what the firm actually knows. And you don’t have the headcount to hire a knowledge manager whose only job is fixing this.

That’s the gap agents fill. Not a new department. A layer of automated work that runs quietly in the background and shows up as saved partner hours and fewer QA surprises.

If you want a broader view of how this plays out across firms structuring their advisory practice around AI rather than around headcount, our advisory guidance walks through the operating model shift in more detail, and our insights library has a few write-ups on how firms sequence these builds without disrupting live engagements.

The Omni Audit, 60 minutes, no deck

We don’t ask you to sit through a sales pitch to figure out if this applies to your firm. We run an Omni Audit. It’s 60 minutes, and you walk away with three specific outputs: a map of where your firm’s manual hours are actually going, a rough dollar estimate of what that’s costing you a year, and a short list of which agent would move that number first. No slide deck, no follow-up sales call unless you ask for one.

For a firm your size, this usually surfaces one or two areas that are costing more than the owner expected, often in the QA and rework layer specifically because it’s the least visible cost on the P&L. You can see Omni for consulting firms before you book anything, or go straight to the audit itself. If you want to get a number on your own firm’s version of this, book a 60-min Omni Audit and we’ll walk through it live.

If you want to try something smaller first

Not every firm is ready to commit to a full audit on the first conversation, and that’s fine. We put together a practical worksheet called Deploy Your First Business Agent that walks you through picking one process, usually proposal drafting or research synthesis, and scoping what a first agent build actually looks like before you spend a dollar. It’s the same framework we use internally, just without us in the room. You can grab the direct download here and work through it with your team this week.

If you’re earlier in your thinking and want to understand what these agents can and can’t do before committing to anything, our guides section has plainer breakdowns of how firms are deploying agents across ops, voice, and app-based workflows, and the blog has more specific breakdowns by function if proposal work isn’t your biggest bottleneck.

The dollar reality

Run the math on your own firm for a minute. Take your average fully loaded partner rate. Multiply it by the hours your senior people spend each month on proposal writing, research they’ve probably done before, and the final QA pass that catches what earlier drafts missed. For most firms in the $1M to $25M range, that number lands somewhere in the $80K to $300K annual range once you’re honest about it, and that’s before counting the deals lost because a proposal went out a day late or a deliverable had an error that made a client nervous.

That’s not a hypothetical. It’s the gap between a firm that treats its own knowledge as an asset and one that rebuilds it every single time. The agents exist to close that gap, and the audit exists to show you exactly how wide it is in your firm specifically.

If you’re ready to see the actual number for your business rather than a category average, book a 60-min Omni Audit and bring your last three proposals. We’ll show you where the hours went and what it would take to get most of them back. Or start with the AI audit for consulting firms and see what it covers before you put anything on the calendar.