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.
There's a startup right now building a product that does what your firm does.
Not a tool to help your firm. A replacement for it.
They have a frontier model, a small sales team, and a pitch deck. They're probably raising $3M, maybe more. And their thesis is straightforward: consulting is about to become software.
I understand why investors love this bet. Y Combinator now lists AI-native agencies in its Requests for Startups, arguing that agencies of the future will look more like software companies, with software margins. The global professional services market is far larger than SaaS. If you can deliver at 20-30% of the cost of a traditional firm and pocket the margin, the math is seductive.
But I've spent the last 18 months doing discovery calls with consulting firms, talking to former partners at MBB shops, boutique founders, and fractional executives building practices around their expertise. And the more I dig into the AI-native consulting firm thesis, the more I think it misunderstands where consulting value actually comes from.
The consultants I talk to didn't start their firms at 25. Most put in 12, 15, sometimes 20 years inside big companies first. They ran departments. They sat in the room when the wrong decision got made. They learned, the hard way, that the client asking for a website usually needs a repositioning. That the sales team asking for more leads often has a pricing problem. That the ops director asking for a process doc sometimes has a hiring problem.
None of that reasoning is available in a public training set.
It lives in their heads. It surfaces in a conversation. It's the thing that makes a client say "you understood us before we understood ourselves." That's not a prompt. It's pattern recognition built across 200 engagements.
One founder I spoke with put it plainly: her value is knowing which questions to ask before the client knows what they need. She's been in enough rooms to see the shape of a problem before it's fully described. A model trained on public data can approximate that. It can't replicate it.
To be fair, AI-native service firms are not fake.
In certain categories, they're formidable. Doc review. Data transformation. Repeatable research tasks. Content at scale. Any work that is fundamentally deterministic, where inputs and outputs can be defined, these firms can deliver faster and cheaper than a traditional firm with headcount.
In our own research, roughly 40% of a consultant's client delivery time goes into data collection and administrative tasks. That's real ground for automation to take.
But here's what gets glossed over in the pitch decks. The work that commands premium fees, the work clients renew and refer, is almost never in that 40%. It's in the judgment layer. The strategic call made with incomplete information. The political navigation inside a client organization. The knowing when to push and when to hold.
A former Oliver Wyman consultant I spoke with described how large consulting firms organize around this. Different practice groups track completely different triggers, use different frameworks, and speak different languages to clients. It's not arbitrary fragmentation. It's because the judgment required to serve a healthcare operator is not the same judgment required to serve a PE-backed software company, even if both are "management consulting" on paper. Harvard research on how consulting firms manage knowledge found the same thing decades ago: the firms selling customized judgment invest in networks of experienced people, not document libraries.
A startup with a model and a sales team can pattern-match on those categories. What it can't do is reason the way someone who has done 200 engagements reasons.
Boutique firms feel the heat from AI-native competitors most in two scenarios.
If you're selling a service that could be described in a clear, repeatable brief with predictable outputs, you're in the AI-native firm's wheelhouse. You're not being disrupted because of AI. You were always doing undifferentiated work.
This is the more interesting problem. A firm might have a founding partner with 20 years of proprietary insight, a way of diagnosing client problems that is genuinely rare, but if that insight lives only in one person's calendar, it doesn't scale. It retires when they do.
The firms most exposed to AI-native competition are not the firms with the deepest expertise. They're the firms that never built systems to capture and deploy that expertise beyond the partners who hold it.
The investment thesis around AI-native firms assumes that buyers care primarily about cost and speed. That's true in some segments and mostly false in others.
A communications firm founder I spoke with told me their primary business development strategy is building deep relationships with clients so they become the first call when something goes wrong. A two-person lobbying firm in Ottawa said they haven't run a single outbound campaign in years. Their pipeline is entirely referral-based. An HR consultancy we've worked with has clients who have been with them for four and twelve years respectively.
Those relationships are not going to a $3M-funded startup with a frontier model. They're maintained because the client trusts that the human on the other end of the phone has seen this before and knows what to do.
The money moves. That part of the AI-native firm thesis is right. Investors have already sold off large consulting stocks on fears that AI will shrink demand. The question is who the money moves toward.
The firms positioned to capture it are the ones who take that 15-year track record and make it systematically accessible. Not just to the partner who built it, but to the whole team. And eventually, to clients directly.
That means getting the playbooks out of people's heads. The frameworks, the decision-making patterns, the "here's what I'd do and why." Captured, organized, and made usable, as an organizational asset rather than an individual one.
We're building Gia around a specific bet: the best-positioned consulting firms in three years won't be the ones that moved fastest to adopt AI tools. They'll be the ones that figured out how to make their proprietary expertise compoundable.
A HubSpot implementation firm we work with now leaves an AI version of their delivery knowledge with clients post-engagement. Clients can ask "why did we build the forecast this way?" at 11pm and get an answer grounded in the firm's actual project data, not a generic LLM response. The firm earns a low-cost monthly retainer for access. The client gets always-on support. The firm's knowledge doesn't walk out the door when an engagement closes.
That's not what an AI-native startup can replicate. They don't have the training data. They never will.
The same tooling is available to everyone. Frontier models, APIs, orchestration layers. The startups have it. Established firms can have it too.
The difference is what you train it on.
I've been wrong about things in this company's history. We've pivoted. We've rebuilt. And we're probably 12 months from knowing if this bet is right.
But the pattern across 900 calls with professional service firms is consistent enough that I'm willing to say it clearly: the threat to boutique consulting is not that AI is smarter than your partners. It's that your partners' knowledge hasn't been made durable.
The recommendation you've given 50 times, the framework you built from watching the same mistake across a dozen clients, the way you diagnose a problem before the client knows what to call it. That's the product.
The question is whether it lives in your head or in your firm.
If you're building systems to capture it, I'd genuinely like to hear what's working. That's the conversation that matters right now.
