Legora moves into the data room, and the legal AI question changes

The interesting thing about Legora's integration with SS&C Intralinks is not that legal AI is now embedded in a virtual data room. It is that distribution, not capability, has just been named as the binding constraint in legal AI — and the company valued at half its main rival's price is the one acting on it.

The announcement. On 5 June, Legora announced that its agentic legal AI now runs inside Intralinks DealCentre AI — the AI layer SS&C bolted onto the dominant VDR (virtual data room, the secure document repository used in M&A). Joint customers can run diligence, drafting, and full document analysis without leaving the data room. Neither side disclosed terms. ¹

Why this matters more than another funding headline. Legora is currently valued at $5.55 billion. Harvey, its main rival, sits at roughly $11 billion. ² The market is pricing the general-purpose platform at twice the vertically-embedded one. This integration is Legora's argument that the market has it backwards.

The thesis hiding inside the press release

The conventional read of legal AI competition is a capability race: whoever builds the smartest model for legal work wins the lawyers. Harvey's strategy fits that frame — a broad legal platform sold firm by firm, partner by partner, expanding feature set against expanding feature set.

Legora's move says something different. It says the binding constraint on legal AI adoption inside investment banks and M&A counsel is not whether the model is good enough. The model is already good enough. The constraint is that the documents live somewhere specific, the workflow runs somewhere specific, the compliance approvals were granted somewhere specific, and that somewhere is the VDR.

Move the AI to the documents. Do not move the documents to the AI. This is not a clever insight. It is the obvious one, which is why it is worth noticing that most of the sector has not done it.

Distribution as the actual moat

The arithmetic of going direct. A legal AI company selling at firm level into M&A practice groups is running a sales cycle measured in quarters, against a procurement function that owes its existence to keeping new vendors out. Multiply that by every Magic Circle firm, every white-shoe firm, every bulge-bracket bank's in-house counsel, and the unit economics of direct sales start to look like a tax on the valuation.

The arithmetic of going through Intralinks. One integration. Access to deal flow that already runs through the platform — the highest-value, highest-stakes legal work in the market. The procurement question collapses, because Intralinks has already cleared procurement. The change-management question collapses, because no individual user is being asked to change anything.

This is the FDE move (functional displacement economics — the idea that AI wins enterprise work by embedding where workflows already run, not by being better in the abstract) made literal. Legora is not displacing the data room. It is displacing the friction that sat between the data room and the AI.

What this is not

It is not a moat. It is a distribution channel, and distribution channels are revocable.

Intralinks has its own AI layer. DealCentre AI exists. SS&C has elected to open it to Legora rather than build the agentic capability itself — which suggests a current judgement that Legora's stack is further along than what SS&C can ship near-term. That judgement is a snapshot. If Legora's product proves the use case at scale, SS&C now has both the customer relationships and the strongest possible reason to replicate the functionality. A successful partnership is the most expensive market research SS&C could buy.

It is also not the whole market. Intralinks is one VDR. Datasite, iDeals, and others hold meaningful share. If Legora's go-to-market depends on VDR partnerships, it has accepted someone else's market structure as its ceiling. Harvey, building firm-by-firm relationships, may end up with stickier ground simply because law firms outlast deal cycles and VDR contracts.

The episodic-versus-continuous problem is sharper still. A VDR relationship is weeks to months long, bounded by the deal. A firm relationship is years. Legora's customers, viewed through the Intralinks lens, are deal teams that disband. Harvey's customers, viewed through the firm lens, are institutions that renew.

The counter-case, taken seriously

The honest version of Harvey's bet is this. Vertically embedding in one workflow wins that workflow until the general-purpose platform catches up. Legal AI's diligence use case is high-volume and economically attractive, but it is not technically exotic. A platform that is good enough at diligence, drafting, research, and twelve other things, sold to the firm rather than to the deal, eventually commoditises the embedded specialist by being present everywhere the embedded specialist is not.

That argument is not wrong. It is a bet on time. Harvey is betting that platform breadth compounds faster than embedded depth. Legora is betting that depth in the right workflow locks in revenue and reference customers before breadth arrives.

The $11B versus $5.55B valuation gap suggests the market currently agrees with Harvey. The Intralinks integration is Legora's argument that the market is mispricing distribution.

What to watch

Three things, in order.

First, whether other VDRs follow. If Datasite announces a comparable Legora integration within twelve months, the playbook is repeatable and Legora has effectively built a horizontal distribution layer across the VDR category. If Datasite announces a partnership with Harvey instead, the market has read the integration as a competitive threat and is choosing sides.

Second, whether SS&C's posture toward Legora changes once usage data accumulates. The most informative signal will be whether DealCentre AI's own roadmap quietly starts to overlap with Legora's feature set.

Third, ARR (annual recurring revenue) disclosure from either company. Both valuations sit on financing-round logic rather than disclosed fundamentals. ² The first credible ARR figure from either Legora or Harvey will tell us which thesis the actual customers are voting for.

Distribution is not a moat. But for the company that gets there first, it is the closest thing to one until the market reprices.

Footnotes

¶