Most "AI" in portfolio tools is a chatbot
Almost every portfolio tool now has an "AI" button — usually a chatbot pinned to a dashboard that hasn't changed. You type a question, it answers from the data already inside the tool, and that's where it ends. Useful, but shallow.
AI-native means something more specific. The data is structured for a machine to read, and your own agents can reach it over an API and an MCP server. That matters if your firm runs AI-first workflows: your team drives Claude, ChatGPT, or an internal agent through real work instead of clicking around a UI.
What "AI-native" actually requires
I use two tests.
- Structured, source-linked data. Every metric has to trace to the document, email, or call it came from. A model that can't see where a number came from can't be trusted with it, and neither can your LPs.
- Read and write access for agents. The tool needs an API and an MCP server so your own agents can pull data and act on it — record a metric, spin up a task, update a company — inside your existing workflow.
A tool that passes both is one your team can build on. A tool that passes neither is a dashboard with a search box.
The tools, ranked
1. Cura
Full disclosure: I built this, so read the rest with that in mind.
Cura turns the raw material around each company — board decks in your inbox, KPI lines in Slack, numbers a founder said on a Granola call, updates they texted you, plus continuous public signal from LinkedIn, X, and news — into structured metrics linked back to their source. The whole record is reachable over a REST API and an MCP server, and Cura's agents both read it and write back to it. Teams point Claude at their Cura workspace to populate custom fields from an LP handout or assign owners from email history, and the results land back in Cura instead of scattering across chat.
It fits solo GPs through institutional VC and growth equity. Pricing starts at $200/month for emerging managers, 90% of customers pay $5,000 to $15,000 a year, and Cura is consistently one of the most affordable options in the category.
Where it's weaker: no peer benchmarking, and companies in deep stealth throw off less public signal, so the founder-reported side carries more weight for them.
2. Standard Metrics
Standard Metrics has moved fast on AI. It offers an AI Analyst for natural-language questions, AI parsing of board decks and financials with a human QA layer, and an MCP server that lets Claude and ChatGPT query your data. Its metrics are traceable to source documents.
The gap is where the data starts and what agents can do with it. Collection runs through a founder portal or connected accounting systems, so it depends on founders reporting each period, and agent access is oriented around querying rather than acting. It's a strong system of record with a capable AI layer on top. See the full Cura vs Standard Metrics comparison.
3. Visible
Visible pairs founder updates with LP reporting, and it fits funds that want a founder-facing portal several investors share. Its AI Inbox parses the financials, decks, and emails founders send, on the Premium plan and up. Its MCP server and API come with the Enterprise plan. It does not monitor public signal, so a quiet company stays quiet in Visible.
How to choose
If you want a benchmark to compare your portfolio against thousands of startups, Standard Metrics leads. If you want a shared founder portal, Visible fits. If you want your own agents to read and act on a structured, source-linked record of everything happening around your companies, that's what Cura is built for.
For the wider category, I also keep an updated guide to the best portfolio monitoring software.