Get 25 battle-tested prompts
Turn real client conversations into Linkedin content that builds authority and attracts the right clients.
Turn real client conversations into Linkedin content that builds authority and attracts the right clients.
Everyone in professional services is having the same conversation right now.
AI is going to eat our margins. Junior work is automating away. Clients will start doing more themselves. The billable hour is dying.
Those fears are legitimate. Accenture's stock has fallen more than 50% this year, including the worst single-day drop in its history, as investors bet that AI will hollow out demand for consulting work. The disruption is real.
But most firms are focused on the wrong threat. The bigger risk is not what AI will take from them. It is what they are already sitting on and doing nothing with.
If you have run a firm for 15 years, you have 15 years of private training data.
Every proposal you wrote. Every client call. Every time you told someone "do not do that, and here is why." The pattern you noticed across twelve different companies all making the same mistake at the same growth stage. The three questions you now ask in the first 20 minutes of any discovery call because you learned the hard way that skipping them costs you three months later.
No public AI model has seen any of it. Your competitors have not either.
That accumulated judgment is the actual reason clients pay your rates. They are not paying for your time. They are paying for the answer shaped by everything you have seen before this moment. It is almost like a pattern library, built privately, available only to you.
And right now, almost nobody treats it like an asset.
It lives in old decks. A few Slack threads. Whoever has been at the firm the longest.
Estimates of how much organizational knowledge is tacit vary widely, but even the conservative numbers are uncomfortable. Panopto's Workplace Knowledge and Productivity Report found that 42% of institutional knowledge is unique to the individual who holds it. For consulting firms, that number probably skews higher. Your methodology is in your head. Your instincts are in your head. Your "we tried that for a client in 2019 and here is what happened" is definitely in your head.
The same research found that employees waste 5.3 hours per week waiting for information from colleagues or recreating knowledge that already exists somewhere. New hires take months to reach full output, mostly because the nuance they need is trapped in senior people who are already stretched thin.
The cost for consulting firms specifically is different. Your people are your product. When one of them leaves, they take an irreplaceable slice of your competitive edge. Gallup estimates voluntary turnover costs U.S. businesses roughly $1 trillion a year, and that figure barely captures the judgment that walks out the door. For a 15-person firm, the math is smaller but the proportional pain is worse. Losing one senior consultant can mean losing years of client-specific context that no handover doc will fully capture.
One pattern I see constantly in conversations with firm owners: a client asks a question three months after a project closes, and nobody left on the team actually knows the answer. The person who built it and made the key decisions is gone. The replacement rebuilds from scratch, in front of the client.
That is not a knowledge management problem. That is a competitive moat slowly draining.
A lot of the advice floating around right now tells consulting firms to "adopt AI tools" and "automate workflows." That is fine as far as it goes. Reducing manual reporting time, drafting SOWs faster, automating scheduling, all of it adds up.
But it also misses the point.
Those tools are available to every firm on the same pricing page. Your competitor can buy the same stack tomorrow. Operational efficiency with generic AI tools is a floor, not a ceiling.
The ceiling is what you do with the knowledge only you have.
One former Oliver Wyman consultant put it clearly in a recent conversation: "If you can say, 'we have 15, 20 years of private training data that no LLM has ever seen,' you are not going to get disrupted by any large language model." That is the actual defensibility story. Not which AI tools you subscribed to. What you built that nobody else can replicate.
An agency founder I spoke with recently illustrated this from the opposite direction. His firm had reduced reporting time from over 100 hours a month to 10 hours across 75 clients by automating the intelligence tasks, the data pulls, the formatting, the slide construction. That freed up his team for the part that actually cannot be automated: the strategic context, the "here is what this data means for your specific situation," the judgment call. The part no model trained on public data can replicate with the same accuracy.
The firms I expect to be fine in five years are not necessarily the ones moving fastest on AI tooling. They are the ones who understand that their historical engagements, decisions, and hard-won learnings are a form of proprietary training data that compounds over time.
For decades, consulting firms have had to pick a lane. Harvard research on how consulting firms manage knowledge found two distinct approaches: codify knowledge into documents and reuse it, or rely on networks of experienced people to share it directly. The best firms committed to one. AI changes that tradeoff, because it can capture the judgment that used to only travel person to person.
What does that look like in practice? A few patterns worth paying attention to:
Getting judgment out of people's heads means documenting not just what was done, but why. The decisions, the supporting logic, whether the outcome was net positive or negative. Most firms document deliverables. Almost none document the reasoning behind them.
Not buried in a Google Drive folder structure nobody uses. Actually queryable. When a junior consultant needs to know how a similar client situation was handled two years ago, they should be able to ask and get a useful answer in seconds, not spend three days excavating old files.
This one is underexplored, and it is probably where the biggest upside sits. Right now, a client hires a firm and gets access to their expertise during scheduled touchpoints: calls, deliverables, check-ins. Between those moments, they are on their own. A marketing agency I spoke with recently is already experimenting with leaving a client-facing AI behind post-engagement on a low monthly retainer, trained on the firm's expertise and the specific project history. The client gets 24/7 access to answers grounded in everything the firm has learned. The firm gets a recurring revenue stream and ongoing visibility into what the client is thinking about next.
That is a fundamentally different value proposition than "we will have a call every two weeks."
The obvious objection: if clients can query your firm's knowledge directly, why do they need you?
It is a reasonable concern and worth taking seriously. But the firms working through it are landing in the same place. The high-touch strategic work is not what gets replaced. It is the friction between touchpoints that gets removed.
A sales advisory firm I have spent time with uses a client-facing AI to grade and coach frontline sales reps daily, based on call transcripts and the firm's methodology. That does not replace their quarterly training kickoffs. It makes the quarterly training stick better, because the reps have been practicing against the firm's IP every week. The firm became more valuable, not less.
The firms worried about cannibalization are often conflating two different things: the commodity parts of what they deliver, and the judgment underneath it. AI can surface the commodity parts faster and cheaper. The judgment is what clients are actually paying for, and making it more accessible does not devalue it. It makes it stickier.
There is a timing dimension to this that does not get discussed enough.
Right now, the firms that start treating their historical engagements as a proprietary data asset have a real advantage. The models they build on top of that data will be better than any generic alternative, because they will be trained on problems that look exactly like their clients' problems.
That advantage compounds. Two years of captured firm judgment is worth more than one year. Five years is worth more than two. A firm that starts now will have a meaningful head start on one that waits until the tools are more obvious.
The firms that wait for AI to "mature" before doing anything with their historical knowledge are not being cautious. They are just falling further behind.
The next version of client delivery in professional services does not look like more software dashboards. It looks like expertise that travels with the engagement.
A client gets access to a firm's 15 years of hard-won patterns, structured and queryable, available when they need it. The firm gets visibility into what the client is thinking about between calls. The relationship gets stickier. The value delivered between touchpoints gets higher. And the firm's accumulated knowledge actually compounds into something that gets harder to replicate the longer they invest in it.
Gia is building for that future. The infrastructure to capture what firms know, how they work, how they make decisions, and eventually make it available to their clients. We are early in this, and so is everyone else. That is exactly why it matters that firms start treating their institutional knowledge as an asset now.
The judgment accumulated over 15 years of client work is not just history. It is the most valuable data your firm has ever produced. The only question is whether it stays locked in people's heads, or becomes something the whole team and eventually your clients can use.
Gia is an AI platform built for professional service firms. We help firms capture their institutional knowledge, make it accessible to their teams, and deploy it to clients. If you are curious what that looks like in practice, we would love to show you.
