Best AI Tools for LP and Fund-of-Funds Due Diligence

The best AI tool for LP and fund-of-funds due diligence in 2026 is Kepler, because it reads every document a GP sends (DDQs, LPAs, side letters, quarterly reports, capital account statements and notices) and turns them into answers an investment committee can use. Look-through exposure arrives broken down by vehicle, terms come back as complete lists across the portfolio, cash flows reconcile, and every number is cited to its statement.

This list is for pensions, endowments, foundations, sovereign funds, fund-of-funds, OCIOs, secondaries buyers and family offices investing in private funds. It covers manager selection and diligence, monitoring after commitment, and the operations work that turns GP documents into data.

The category moved fast in 2026. MSCI bought the AI diligence startup Vantager in March, Nasdaq completed its purchase of Dasseti in September, and Bloomberg completed its acquisition of Canoe Intelligence in October. Cobalt LP, a name on many older lists, has been part of Hamilton Lane’s technology business since 2020.

How we picked

• Document coverage: does it read every page of every GP document, including scans, or collect files and extract a fixed set of fields?
• Cross-fund answers: can it answer questions that span funds and documents, such as exposure to one company or every key-person clause?
• Numbers that reconcile: are figures traced to their statement, and do calls, distributions and NAV tie out?
• Diligence workflow: DDQs, manager comparison and track-record analysis.
• Controls: permissions by fund and team, audit logs, and a stated position on whether your documents train anyone’s models.

Kepler is our product and appears first, labeled as such. Competitor details come from each vendor’s site, press releases and news reports, checked in November 2026. None of the specialist vendors here publishes pricing.

Quick comparison

Tool	Best for	Documents it handles	AI features
Kepler	One model of your whole program, every number cited	Every GP document, plus your CRM, accounting and inbox	Ontology, full-document reading, numbers computed in code, agents
Dasseti (Nasdaq)	DDQ-driven manager diligence	DDQs, ADV filings, policies, manager reports	Sidekick assistant: extraction, gap and inconsistency flags
Chronograph	LP portfolio monitoring	Quarterly reports, statements, notices, LPAs	ChronoAI, semantic search, Claude plugin with look-through
Canoe Intelligence (Bloomberg)	Document collection and extraction at scale	Capital account statements, reports, notices	AI extraction across 44,000+ funds
MSCI Private Capital Solutions	Benchmarks, manager screening, data	GP data-room materials, GP and administrator reporting	Diligence Platform, AskPCI, MCP connector
Arch	Family offices and allocators that want collection plus look-through	Calls, valuations, distributions, K-1s, offering documents	Plain-English questions, flags linked to source
Addepar	Wealth managers and family offices with alternatives	Capital calls, statements, distribution notices	AI extraction with review, Addison assistant
Hebbia	Running one question set across a stack of fund documents	Whatever you load or connect	Matrix agents with citations
Claude / ChatGPT for Financial Services	A frontier assistant with LP data connectors	Uploads and connected sources	MSCI, Chronograph and Addepar connectors

The best AI tools for LP and fund-of-funds due diligence

1. Kepler (our pick)

We build Kepler.

Best for: LPs, fund-of-funds, endowments, pensions, OCIOs and family offices that hold dozens or hundreds of fund interests and need answers that cut across every document their GPs send.

An allocator’s data arrives as documents from dozens of GPs, each with its own layout. Most software in this category either collects those files or pulls a fixed set of fields into a database. The questions an investment committee asks cut across all of them (“What is our look-through exposure to this company?” or “Which funds have a key-person clause naming this partner?”), and they still end up as a spreadsheet project.

Our engineers build one living model of your program. Every fund, vehicle, GP, key person, commitment, portfolio company and co-investment resolves to one record in your firm’s ontology, so a company held through four funds under four spellings is one entity. Extraction models we train read every page of every document at ingest, scanned statements included, and keep each fact beside its passage. In a test we published, our reader went through a 28 MB fund report end to end in 136 seconds and recorded 83,204 facts, 45,199 entities and 9,042 relationships (Stop Paying to Forget). Your definitions of net IRR, TVPI, DPI, unfunded commitment and exposure live in the model and apply every time. We route each job to the right model: small models we train read the statements and notices, frontier models answer the question and draft the board report, and we change models whenever a better one ships.

With that in place, the hard questions take one step. Look-through exposure arrives as one total, split by fund, co-investment and account, with each input cited to its statement and the arithmetic done in code. A key-person question returns every agreement that has one, with the clause text from each LPA and any side letter that changes it. Capital calls and distributions reconcile against each capital account statement, with breaks flagged (how to process capital calls with AI). Agents read every notice and report the day it lands, acting as the person who set them up, and alert you with the exact quote. The quarterly board report and the re-up memo draft in your format, from the templates and manager scorecards your team saves once. Our forward-deployed engineers built this layer before at Palantir; they connect your portals, inbox and data stores and have your ontology running in the first week.

Why it’s our pick for LPs and fund-of-funds

• Look-through exposure across every fund, co-investment and separate account: one number, broken down by vehicle, every input cited.
• Complete lists from every LPA and side letter: key-person provisions, MFN rights, no-fault removal, fee offsets.
• Capital calls, distributions and NAV reconciled to each statement, with breaks flagged.
• Manager diligence across years of quarterly letters and DDQs, every claim linked to its source.
• Kepler in Excel, PowerPoint, Word, Claude, ChatGPT and Slack, and in your own systems through the API.
• Permissions by fund and by team, every query logged, and data isolated per firm. We hold SOC 2 Type II (security).

Questions allocators ask Kepler

• “What’s our look-through exposure to this company across every vehicle, as of the latest statements?”
• “List every fund agreement with a key-person clause, the people it names, and what happens to the investment period if it triggers.”
• “Reconcile Q3 capital calls and distributions against each capital account statement and show the breaks.”

“Our analysts used to spend more time verifying AI outputs than generating them. With Kepler, every number is already traced to source.” (Senior Managing Director, Top-10 Global Asset Manager)

Not the right fit if: you hold a handful of fund interests that a spreadsheet keeps up with, or you’re looking for a fund administrator’s books and records.

Book a demo to see Kepler on fund documents like yours, or try it on public companies at app.kepler.ai.

2. Dasseti (Nasdaq)

Best for: allocators and consultants that run manager diligence through DDQs.

Nasdaq completed its acquisition of Dasseti in September 2026 and is folding it into Nasdaq eVestment. Dasseti COLLECT is due diligence and monitoring software for institutional investors, and ENGAGE handles the manager side of RFPs and DDQs. Its AI assistant, Sidekick, extracts information from ADV filings, policies and manager reports and maps it to DDQ questions, highlights missing information and inconsistencies, scores sentiment to flag operational and investment risks, compares managers’ responses and suggests follow-up questions. Smart Docs pre-fills DDQs and manager profiles from ADV filings, policies and reports, and Knowledge Base queries the firm’s internal and connected documents. Sidekick runs on Azure OpenAI.

How Kepler differs: Dasseti structures the questionnaire process. We read everything around it, from years of quarterly letters to every LPA and side letter, so Kepler can check what a manager says in the DDQ against what its reports show.

3. Chronograph

Best for: LPs that want a dedicated monitoring platform with AI across their fund documents.

Chronograph builds portfolio monitoring for LPs and GPs. ChronoAI works across quarterly reports, capital account statements, calls and distributions, LPAs and AGM materials, and Semantic Search (March 2026) searches an investor’s whole document portfolio in plain language. In October 2026 it released an LP plugin for Claude covering cash-flow forecasting with the Takahashi-Alexander model, commitment pacing, look-through exposure by sector, geography, vintage and currency, and manager-meeting prep. Its connector is live in Claude, ChatGPT and Perplexity.

How Kepler differs: Chronograph organizes and monitors your fund data. We build one model of your whole firm from every document, and Kepler computes on your definitions in code and runs agents that act on what arrives. See Kepler vs Chronograph.

4. Canoe Intelligence (Bloomberg)

Best for: large allocators and wealth platforms that need GP documents collected and extracted at scale.

Canoe uses AI to extract data from capital account statements, performance reports, notices and spreadsheets, across more than 44,000 funds. Bloomberg agreed to buy Canoe in July 2026 and completed the deal on October 1; terms were not disclosed. For joint clients, Canoe’s validated fund data flows into Bloomberg’s PORT Enterprise for a total-portfolio view.

How Kepler differs: Canoe turns documents into a feed of standard fields. We read every page, including the terms and commentary a field list leaves out, and Kepler answers questions across the whole program with every number cited.

5. MSCI Private Capital Solutions

Best for: allocators that want benchmarks, manager screening and private-asset data from one provider.

MSCI’s private capital business (formerly Burgiss) runs Private i, Total Plan Manager and Private Capital Intel. Its AI page lists AskPCI inside Private Capital Intel, agentic collection and extraction of GP and administrator reporting, a Private Assets MCP connector available in Claude and ChatGPT, and the MSCI Diligence Platform, which extracts data from unstructured GP materials and produces AI manager-screening reports. That platform came from MSCI’s March 2026 acquisition of Vantager, an AI tool built to work across GP data rooms.

How Kepler differs: MSCI’s strength is its cross-market data and benchmarks. Kepler’s is your own program: your funds, your terms and your definitions, resolved into one model and computed in code.

6. Arch

Best for: family offices and allocators that want document collection and look-through in one product.

Arch collects capital calls, valuations, distributions and K-1s across funds with a human in the loop, flags changed wire instructions, shows look-through company and sector exposure adjusted to your ownership share, and answers plain-English questions. In September 2026 it extended its AI to pre-investment diligence, and its investment research product pulls deal terms from offering documents and PPMs, linking each flag to its source.

How Kepler differs: Arch is built around collection and reporting workflows. Kepler is built around a model of your firm that every question, agent and report draws on, with your own definitions applied in code.

7. Addepar

Best for: wealth managers and multi-family offices with a large alternatives book.

Addepar is a wealth and portfolio reporting platform with a growing set of tools for private funds. Alts Data Management uses AI to extract and map data from capital calls, statements and distribution notices, with human review, and Navigator forecasts calls, distributions and NAV and models commitment pacing. At its May 2026 conference Addepar announced expanded private fund look-through and capital activity dashboards. Addepar launched its Addison assistant earlier in 2026 and offers an MCP connector for Claude.

How Kepler differs: Addepar reports on the portfolio. Kepler covers the research and diligence side: every fund document read in full, terms across every agreement, and manager analysis with every claim sourced.

8. Hebbia

Best for: teams that want to run one set of questions across a stack of PPMs, LPAs and DDQs.

Hebbia has no LP-specific page, but its Matrix runs the same questions across any set of documents, with agents filling each cell and citations back to the source. Its integrations include Preqin fund and LP data, PitchBook, Box, SharePoint, Egnyte, Datasite and Intralinks.

How Kepler differs: Matrix works on the documents you select for each question. We read every document once and resolve funds, managers and companies into your ontology, so Kepler returns complete answers across the program. See Hebbia alternatives.

9. Claude and ChatGPT for Financial Services

Best for: teams that want a frontier assistant connected to the LP data they already have.

Claude for Financial Services connects to PitchBook, Morningstar, Box and others, and LP-specific data reaches Claude through the MSCI, Chronograph and Addepar connectors. On the OpenAI side, ChatGPT for Financial Services (September 2026) bundles PitchBook Essentials and more than 50 connectors, and both MSCI and Chronograph say their connectors work there. Kepler is available inside both assistants.

How Kepler differs: by itself, the assistant sees only what you upload or what a connector hands back, and it writes figures itself. With Kepler in the same window, every fund document is already read and resolved, and every number is computed and cited.

Also worth a look

• Cobalt LP: portfolio monitoring and benchmarking for LPs, now part of Hamilton Lane’s technology offering.
• Allocator: data management for LPs that collects GP and hedge fund reports and extracts data from quarterly reports and ILPA templates, validated by people.

How to choose

1. Separate the jobs. Pre-commitment diligence (DDQs, data rooms, track records), post-commitment monitoring (reports, statements, exposure) and operations (notices, reconciliation, K-1s) can be different tools or one.
2. Ask the cross-fund question. “What’s our exposure to this company across every vehicle?” shows quickly whether a tool understands your program or only individual documents.
3. Test with a messy GP. Use your worst-formatted manager’s scanned statements as the test set.
4. Count the list. Ask for every fund with a key-person clause and compare against your legal team’s tracker.
5. Tie out a quarter. Have the tool roll forward one fund’s capital account from the notices and see whether it matches the statement.
6. Check permissions. Fund documents are confidential under the LPA, so the tool should mirror who can see what and log every access.

For the document-by-document detail, read how LPs and fund-of-funds use AI on fund documents. For why entity resolution matters so much here, see what an ontology is, and for wider tools, the best AI tools for asset managers. Allocators that also co-invest should see the best AI tools for private equity due diligence.

Try Kepler at app.kepler.ai, or book a demo and we’ll show look-through exposure on a program like yours.

Frequently asked questions

What is the best AI tool for LP due diligence?

Kepler is our pick for allocators that want one model of their whole program. Look-through exposure arrives as one number broken down by vehicle, terms come back as complete lists across every LPA and side letter, capital calls and distributions reconcile against each statement, and every number is computed in code and cited. We build Kepler.

Can AI calculate look-through exposure across our funds?

Yes, if it can read every fund's schedule of investments, match the same company across different GPs' spellings, and apply your ownership share in each vehicle. Chronograph, Addepar and Arch all offer look-through views. Kepler returns look-through exposure as one number broken down by vehicle, with every input cited to the statement it came from and the arithmetic done in code.

What is the difference between document collection tools and AI diligence tools for LPs?

Document collection tools such as Canoe and Arch gather GP documents from portals and inboxes and extract a set of fields, such as calls, distributions and NAV, often with human review. Diligence tools such as Dasseti and MSCI's diligence platform focus on DDQs and manager screening before a commitment. Kepler covers both jobs on one model of your program, so the terms you diligenced are the terms you monitor.

Can ChatGPT or Claude analyze an LPA or side letter?

They can summarize one and answer questions about its terms, and both now reach LP data through connectors from MSCI, Chronograph and others. On its own, a frontier model reads what you upload or retrieve, so a question across 200 LPAs returns its closest matches. Kepler runs inside both and returns the complete list, with each clause quoted from its source.

What is the ILPA DDQ?

The ILPA Due Diligence Questionnaire is a standard set of questions LPs send to GPs before committing to a fund, covering the firm, team, strategy, track record, operations, ESG and more. ILPA introduced it in 2013 and last updated it in 2021 as DDQ 2.0, which added sections on co-investments, GP-led secondaries, credit facilities and data security.