Datasite unifies AI across M&A lifecycle with $500m CapVest backing
Datasite, the M&A data-room and intelligence platform backed by private equity firm CapVest Partners, has launched two products that extend artificial intelligence across the full deal cycle: Grata Intelligence, an agentic workspace for target discovery and deal execution, and Datasite Connect, which lets enterprise teams pipe approved third-party AI assistants into live transaction documents without loosening document-level permissions. The announcements advance a $500 million investment commitment from CapVest and integrate two acquisitions Datasite completed in 2025, private-market intelligence provider Grata and agentic AI developer Blueflame AI.
The practical change for dealmakers is meaningful. Previously, a corporate development team would move between separate tools to screen targets, build a shortlist, and then conduct document-heavy diligence inside a secure data room. Grata Intelligence collapses those steps into a single permissioned workspace; Datasite Connect extends the same logic to teams that prefer their own AI tooling. Critically, the company says access controls are inherited from existing project permissions, so a user's AI assistant can only surface content that user was already authorised to see.
The tension between AI adoption and data control
That permissions architecture reflects a structural tension now shaping the entire private-markets technology stack. As AI adoption accelerates across investment banking, private equity and corporate M&A, deal teams are caught between pressure to deploy AI widely and their fiduciary and legal obligations to ring-fence sensitive transaction data. A March 2026 survey cited by Datasite, conducted by Accenture, found that senior dealmakers and C-suite executives expect agentic AI maturity in post-deal integration and value capture to grow by 72% in the near term.
"Deal teams shouldn't have to change how they work to benefit from AI," said Andrew Bocskocsky, Chief Product Officer at Datasite. "Customers can use integrated intelligence within Grata or connect their own approved AI tools to active transactions, choosing the option that works best for them."
Datasite's positioning is notable for its breadth: the company says it is already trusted by all top 20 global financial advisory firms and 74 of the top 100 global law firms, and that its platform carries visibility into transactions up to nine months before public announcement. That data moat is now the substrate on which the new AI layer sits, and it is a credible differentiator in a segment where incumbents have historically competed on security and workflow depth rather than intelligence.
Cross-sector implications for AI-native financial infrastructure
The Datasite build-out is worth reading as a capital-allocation signal beyond M&A technology itself. CapVest's $500 million commitment, deployed through a combination of organic product investment and acquisitions completed inside a calendar year, follows a wider pattern of private equity backing AI-native infrastructure plays at the intersection of financial services and data. The hypothesis being tested is that the next generation of deal infrastructure will not be won by point solutions (a stand-alone data room here, a target-screening tool there) but by platforms that own the full deal lifecycle and can therefore charge a premium for integrated, permissioned AI.
That logic has direct read-across for enterprise software broadly. The same architectural question, how to deploy agentic AI without dissolving access controls, is being confronted simultaneously in healthcare records management, legal discovery, and defence procurement, wherever large volumes of sensitive documents must be queried at speed. Whichever sector resolves the permissions problem most elegantly will set a template others adopt. Datasite is making a well-capitalised claim that M&A is where that template gets written first.
Grata Intelligence is available immediately; Datasite Connect is scheduled for release on 4 November 2026.