How We Built an AI Prompt to Help Investors Perform Due Diligence on Private Funds
When Microsoft Copilot became available last year, we saw an opportunity—not just to experiment with AI, but to integrate it meaningfully into our workflows. Since then, we’ve embedded AI into our processes across the company, from underwriting and loan servicing to investor communications and compliance reviews.
But one area where we felt AI could be a valuable tool is in due diligence.
There’s been a lot of discussion in one of the investor forums we’re part of about using AI to perform due diligence on private investment funds. We’ve read the commentary, and we agree: AI has the potential to help investors ask sharper questions and make more informed decisions.
So we decided to take it a step further.
🛠️ We Built the Prompt Ourselves
We didn’t just test someone else’s idea—we built the prompt from scratch, leveraging my multi-decade institutional-level experience in performing due diligence.
Our goal was to create a structured, investor-focused tool that could:
Review fund documents objectively
Benchmark fund terms against industry standards
Flag areas of uncertainty or missing information
Analyze from multiple perspectives: legal, investor, and lender
And to see if it worked, we tested it on our own fund—the Kirkland Income Fund I.
🔍 What the Prompt Does
The prompt guides AI to perform a comprehensive review of fund documents, covering:
Capital call provisions
Valuation practices
Fiduciary duties
Fee structures
Liquidity terms
Use of leverage
Governance and compliance
It also includes a benchmarking table that compares fund metrics (like target returns, GP commitment, and default rates) against authoritative sources like PitchBook, ILPA, and McKinsey.
And finally, it generates an Uncertainty Map—a list of vague, missing, or conflicting information in the documents, with suggested follow-up questions for fund managers.
If you feel the prompt needs to be adjusted to better suit your preferences or the type of fund you’re reviewing, feel free to edit or optimize it. We know it’s not perfect and it’s designed to be flexible.
🧪 Testing It on Our Own Fund
We used the Kirkland Income Fund I as the test case. Why? Because we wanted to see how the prompt performed on real documents—and we were confident enough in our own transparency to put it to the test.
The results were impressive.
The AI flagged strengths like:
No carried interest
High GP commitment (6.71% of AUM)
Monthly NAV updates
Strong collateral enforcement
It also identified areas for follow-up:
Lack of a formal compliance manual
Internal valuation oversight
Early withdrawal hardship criteria
And while some of the AI’s interpretations needed human clarification (like its assumptions about UBTI exposure, which don’t apply to our REIT structure), the overall output was detailed, structured, and useful enough for a potential investor to craft more educated questions for us, the fund manager.
🤝 Why This Matters for Investors
Due diligence is hard. Documents are dense. Terms are complex. And investors often don’t know what to ask.
This prompt helps bridge that gap.
It doesn’t replace human judgment—but it gives investors a start. It helps them spot red flags, ask better questions, and compare fund terms against industry norms.
And because we tested it on our own fund, you can see exactly how it works. We’ve even added a commentary on what areas it did well—and where it missed the mark.
📥 Download the Prompt + Case Study
We’re making the full prompt available as a free download, along with the unedited AI output from our Kirkland Income Fund test.
👉 Download the Prompt + Case Study
Whether you’re an individual investor, a family office, or an advisor, this tool can help you take the first step toward smarter, AI-assisted due diligence. And remember—it’s not perfect so you can tailor the prompt to suit your own review style or fund type.
📚 Explore More
Visit our Investor Education Page for more content on due diligence and alternative investments.