.svg)
.svg)
Why this matters
AI assistants are only as useful as the data they can reach. Until now, that meant copying figures into a chat window, working from a stale export, or describing a situation from memory and hoping the AI could do something useful with it. That's a workaround, not a workflow.
Atominvest's MCP integration changes this. Your portfolio data, fund metrics, investor records, and commitment detail are now directly accessible to the AI tools your team already uses - live, in context, and without leaving the assistant.
What is MCP?
Model Context Protocol (MCP) is an open standard that lets AI assistants connect securely to external software. Atominvest's MCP server acts as a secure bridge between your preferred AI assistant and the Atominvest platform, so you can ask natural-language questions about your portfolio companies, funds, investors, KPIs, and commitments - and get answers instantly. Nothing is stored, nothing is cached, and your data never leaves your control.
Built for trust:
- Read-only: MCP extracts and analyzes your data without changing it.
- Your data stays yours: your data is never stored on the MCP server.
- Permission-aware: access is governed by your Atominvest credentials and scopes.
- Tenant-isolated: your API key routes requests exclusively to your environment
For Portfolio Management teams
Turn your governed portfolio dataset into an on-demand analyst. Ask questions across funds, assets, and KPIs and get answers in seconds.
What you can do today:
- Query portfolio, fund, and KPI data in natural language across your whole book
- Pull asset metric data on demand, optimized for fast, clean AI analysis
- Retrieve investment performance, including returns and cashflows, without exporting or navigating between pages
- Access individual metric definitions and asset-level transaction detail directly from your AI assistant
Example queries:
"Summarize revenue and EBITDA trends across my portfolio this quarter."
"Which assets have the highest cash burn right now?"
"Show me the latest returns.”
For Investor Management & Investor Relations teams
IR teams spend significant time on questions that should take seconds. MCP gives your team instant answers on investors, funds, and commitments without digging through the platform or chasing someone for a spreadsheet - so the time that used to go on data retrieval goes on the relationships instead.
What you can do today:
- Query investor, fund, and commitment data in natural language
- Run investor and commitment analysis directly from your AI assistant
- Speed up reporting prep and ad-hoc investor questions
Example queries:
"How much has each LP committed across our funds?"
"Summarize commitment activity for this fund."
"Which investors are in more than one vehicle?"
Your data, your controls - not ours
The MCP doesn't sit between you and your fund operating system. It extends access to it, on your terms. Authentication, permissions, and data scoping are all inherited from your existing Atominvest configuration - there is no separate permissions model to manage, no shadow data layer, and no access that exists outside what you've already defined.
As AI becomes a more central part of how IR and portfolio teams work, that control layer matters more, not less. Atominvest MCP is designed so that the convenience of AI doesn't come at the cost of governance.
Getting started
Connecting takes a few minutes:
- In Atominvest, go to Settings → API & Webhooks → MCP Credentials and generate a new MCP key
- In your AI assistant, add Atominvest as a custom connector (Claude) or custom app (ChatGPT)
- Authenticate with your MCP key and start asking questions
What's next
The current release is read-only and focused on retrieval, which is a deliberate starting point. Write capabilities, agent-driven actions, and deeper workflow integrations are on the near-term roadmap. How teams use this version will shape what we build next.
Want to see it in action? Get in touch for a live demo.
.jpeg)
