Merging agentic AI, MCP, and human-led insight in Private Equity
This report was published in collaboration with Actum Group.
The private equity (PE) industry is crossing a structural Rubicon. For years, “digitization” meant moving files to the cloud and using basic screening tools. Today, the frontier has moved from simple data access to autonomous reasoning.
As firms like Ardian, EQT, Brookfield, and Brookstreet Equity Partners redefine their operational models, a new architecture is emerging—one where agentic AI, the model context protocol (MCP), and human-led insights converge to create a permanent information edge.
The structural shift: Bridging deal teams and operations
A significant trend observed at the recent Actum event is the collapsing of silos between investment (deal) teams and value creation (operating) teams. Historically, these groups operated in sequence: the deal team bought the asset, and the operating team moved in post-close.
Leading firms are now embedding operating partners from the very start of the sourcing process through to the final exit. This lifecycle involvement requires a unified intelligence layer. If an operating partner has deep expertise in supply chain resilience, that insight needs to be available to the deal team during the initial 48-hour bid window. AI is the bridge that makes this cross-pollination scalable.
Tristan Tully, Managing Partner, Head of Private Equity, Europe, Brookfield says that AI is increasingly embedded across the investment lifecycle, but its real value is realized when paired with experienced operators who can translate insight into execution.
“At Brookfield, our integrated investment and operating teams deploy AI across the full value chain - enhancing underwriting and driving operational transformation in industrial businesses, from demand planning and supply chains to plant-level productivity. We see AI not as a separate strategy, but as a force multiplier within our value creation playbook - enabling more resilient businesses and delivering measurable, system-wide improvements in performance,” he adds.
From chatbots to agentic workflows
While many firms began their AI journey with basic “chatbots” like ChatGPT or Claude, the industry is rapidly moving toward agentic AI. Unlike a chatbot that waits for a prompt, an AI agent is a “digital associate” capable of planning and executing multi-step tasks.
In a modern PE workflow, an analyst can task an agent with “validating the market share claims of Target X.” The agent doesn’t just search the web; it deconstructs the request, identifies the necessary data sources, and autonomously synthesizes a report for the investment committee (IC). At firms like Ardian, this is becoming institutionalized: every investment memo is now required to feature a front-page analysis of how AI will disrupt or enhance the specific investment opportunity.
The MCP connector: solving the data silo problem
The primary bottleneck for agentic AI has always been “the silo.” AI agents are often disconnected from the proprietary, high-value data trapped behind APIs or in locked repositories.
Omiros Sarikas, Managing Partner at Brookstreet Equity Partners, says that AI is unquestionably reshaping private equity, but much of the industry conversation is focused on the surface rather than the foundation. “Access to AI tools is no longer scarce. What differentiates firms today is not who licenses the most sophisticated model, but who has built the discipline, structure and integrity of data required to make those models meaningful,” he says.
The Model Context Protocol (MCP) is the technical catalyst solving this. MCP acts as a universal, secure “pipe” that allows an AI agent to plug into external data ecosystems as if they were part of its own native memory.
The Third Bridge MCP integration
A primary example of this is the Third Bridge MCP. By integrating this protocol, a firm’s internal AI agent gains seamless access to over 100,000 verified expert interview transcripts.
- Grounded truth: Instead of the AI “hallucinating” market trends based on public noise, it is grounded in the “ground truth” of human-to-human interviews.
- Native connectivity: Analysts no longer need to copy-paste data; the agent “reads” the Third Bridge Library in real time to stress-test management’s assumptions.
- Multi-LLM flexibility: Many firms are building front-end systems that toggle between different AI models. MCP provides the standardized plumbing that allows these models to access the same high-quality data regardless of which “brain” is being used.
The human element: Judgement over automation
Despite the move toward automation, the consensus among industry leaders is clear: AI will not replace the analyst; it will replace the “drudge work.”
The role of the PE professional is shifting from a “data gatherer” to a “system orchestrator” or agent architect. This requires a shift in mindset—moving away from rigid KPIs for AI adoption and toward fostering natural curiosity.
The human-in-the-loop remains the most critical component for several reasons:
- The “weighting” of truth: An AI can summarize 50 transcripts, but a human analyst must decide to weight the insights of a former CEO more heavily than a mid-level manager.
- Nuance and negotiation: AI cannot navigate the cultural fit of a founder-led business or the delicate nuances of a complex negotiation.
- Change management: Successful AI integration requires transformation and influence skills to guide a workforce through behavioral shifts.
The real edge: The augmented PE firm
By adopting an infrastructure where agentic AI connects to human-led insights via the MCP connector, PE leaders ensure that their teams spend less time “finding” the news and more time “making” the news. This is the future of the augmented investment lifecycle: a process where technology handles the scale, while humans provide the vision and judgment.
“The firms and PE/VC fund managers that will benefit most from AI are not those that deploy the greatest number of tools, but those that institutionalize rigorous data governance and coherent information architecture. A unified data spine connecting origination insights, diligence analytics, capital structure modeling and portfolio performance is essential. Without that integration, AI remains cosmetic,” concludes Sarikas.
