There is a structural gap in the AI tooling market for M&A that nobody at the top tier has any incentive to fix. The companies building the most sophisticated AI for finance (Rogo, Hebbia, Harvey, Legora) are selling to the institutions that least need help. Meanwhile, the boutique advisory tier that does 80% of the world’s mid-market deals has been quietly waiting for someone to notice it exists.
I want to be clear before I go further: this is not a criticism of those companies. They are building excellent products. Rogo just raised a $160M Series D led by Kleiner Perkins in April 2026, bringing total funding to more than $300M and serving 35,000+ professionals at 250+ institutions including Rothschild, Lazard, Moelis, Jefferies, and Nomura. Hebbia raised $130M at a $700M valuation in 2024 and now processes more than a billion pages a year, serving BlackRock, KKR, Carlyle, Centerview, and reportedly 40% of the world’s largest asset managers by AUM. These are real companies with real customers and real product depth.
But there is a tier they cannot serve, even if they wanted to. And that tier is where most of the actual work in M&A gets done.
What Rogo and Hebbia actually do
Rogo positions itself as Wall Street’s first AI analyst. The product integrates into Excel, PowerPoint, Word, and the firm’s data warehouse. It runs a chat-first interface where bankers can prompt it to build pitch material, run precedent transaction analyses, draft investment committee memos, and increasingly, generate full IM and fairness opinion content. Their August 2025 upgrade to GPT-5 specifically expanded the product into associate and VP-level workflows.
Hebbia is more horizontal, built around their Matrix platform, which uses a multi-agent architecture to perform deep document analysis across very large data sets. Investment bankers use it for due diligence, earnings call synthesis, contract review, and memo drafting. Their 2025 acquisition of FlashDocs added automated deck and memo generation on top of the retrieval layer.
Both are powerful. Both are deeply integrated into the workflows of the institutions that buy them. And both are structurally locked into a deployment model that excludes most of the M&A industry.
Why boutiques cannot buy these tools
Enterprise AI requires enterprise sales motions. That is not a moral failing; it is what the unit economics demand. The CAC of selling a $200K+ annual contract with custom deployment, dedicated AI strategists, and white-glove implementation is too high to support smaller customers at any margin. The customer also doesn’t fit the pattern: a 15-person boutique doing 12 deals a year has no procurement department, no security committee, no enterprise IT, and no budget line item that looks anything like what Rogo or Hebbia is selling.
The result is that the M&A advisory industry has bifurcated. The top tier (the bulge brackets, the elite boutiques like Centerview, the global asset managers) is getting access to AI tooling that compresses their workflows by 30-40 hours per deal, with Hebbia’s own published numbers showing exactly that range. The bottom tier (the 5,000+ lower and mid-market boutiques that run 8-30 mandates a year) is still operating with the same toolkit they had in 2018: Microsoft Office, a data room subscription, and analyst hours.
This is not stable. The competitive advantage that AI gives the top tier widens every quarter that the bottom tier goes without equivalent tooling. Mid-market mandates that used to flow naturally to boutiques will start migrating upmarket as larger firms expand into smaller deal sizes with AI-enabled cost structures.
The layer that has to exist
If you take the structural argument seriously, the question becomes: what does an AI tool for the boutique tier actually look like? It looks different from Rogo and Hebbia in three specific ways.
First, the deployment has to be self-serve or close to it. A boutique cannot absorb a 6-month implementation. The product has to be usable within 48 hours of signup, with the firm’s templates and tone built in within a week.
Second, the pricing has to fit the economics of a firm doing 8-30 deals a year. That probably means per-mandate or per-deliverable pricing, not enterprise seat licensing at $20-30K per seat per year. The math has to work for a firm where the average mandate fee is $200K-$2M, not $20M.
Third, the product has to be opinionated about the workflow. Bulge brackets and elite boutiques have their own internal frameworks for how an IM should be structured, what an equity story looks like, how the financial section ties back to the investment thesis. Smaller boutiques often do not. They need the tool to bring the framework with it, drawn from market patterns rather than built from scratch.
What this means for boutique partners
If you run a boutique, the practical implication is that you have options now that did not exist 18 months ago. The same wave of AI capability that is reshaping the top tier is starting to reach you, just through a different distribution model. The tools won’t look like Rogo. They will look like vertical software built for your specific workflow, sold at prices that fit your P&L, and deployed in weeks rather than quarters.
The risk is waiting too long to evaluate the new layer because it doesn’t show up in the same trade press that covers Rogo’s Series D. Most of the relevant tools at the boutique tier are not yet at the scale where they get written about in Bloomberg. They get found through peer referrals, through analysts who have used them at previous firms, and through founders like me who are building them and writing about them.
I’d take that seriously. The top tier is not going to wait for the bottom tier to catch up. And the boutique tier doesn’t have to wait for an enterprise tool to come downmarket; the right product for that segment looks structurally different anyway.