Questions LPs ask GPs about AI in 2026, and how to answer them
LPs ask GPs about AI in four areas: where AI is deployed across the portfolio and what it has changed, how AI risk is governed, how the firm itself uses AI, and whether AI creates new revenue or threatens existing businesses. The best answers are specific, evidence-backed and staged, and describe new revenue such as data licensing without projecting proceeds.
What LPs are actually asking about AI
LPs ask about AI to test three things: whether the GP's value creation is real, whether portfolio risk is controlled, and whether the GP is ahead of the change or behind it. The questions arrive in annual meetings, quarterly calls, LPAC sessions, fundraising due diligence questionnaires and re-up conversations, and they get sharper each cycle.
The stakes are higher because value creation is now the main story LPs underwrite. McKinsey's Global Private Markets Report 2026 reports that 53 percent of 300 surveyed LPs ranked a GP's value creation strategy among their top five manager selection metrics. AI is one of the first places LPs now look for evidence that the strategy is more than a slide.
The questions, what they test, and how to answer
Group the questions into four areas, adoption, governance, firm practice and revenue, and prepare one honest answer for each with evidence you can show.
| LP question | What it is really testing | A strong answer includes | Avoid |
|---|---|---|---|
| Where is AI deployed across the portfolio? | Breadth versus anecdote | A count of companies by stage: exploring, piloting, in production | One flagship case presented as the norm |
| What has AI changed in the numbers? | Attribution discipline | Specific KPIs moved, such as cycle time or cost per ticket, against a baseline | EBITDA claims with no baseline |
| How do you govern AI risk in portfolio companies? | Downside control | A policy, an owner per company, vendor review and rules on outbound data | Answers that rest only on management's word |
| How does the firm use AI in its own work? | Operating maturity | Uses in sourcing, diligence and reporting, with human review | Claims of automated investment decisions |
| Which holdings does AI threaten? | Honesty about exposure | A documented exposure review and the actions taken | Saying none are exposed |
| Is AI creating new revenue? | Upside without hype | Initiatives by stage, with no projected proceeds until contracted | Pipeline presented as revenue |
| How does AI change your exit story? | Exit readiness | What buyers will test: data, systems and adoption evidence | Valuation uplift claims |
Preparing the governance answer
Governance is where LPs push hardest because it is where losses come from: confidential data pasted into public tools, vendor contracts that let a provider train on a company's records, or customer data used in ways a privacy policy never promised.
A credible answer names the policy, the owner and the control. The strongest version shows that each portfolio company has an AI policy covering approved tools, data classification and an outbound data clause: no company data leaves under any agreement, including a data license, without a documented rights review and approval by the CEO or board. The AI governance policy template includes that clause and can be adapted company by company.
Preparing the new revenue answer
Expect a question on whether AI is a source of revenue as well as savings. One answer some GPs can give is data licensing: portfolio companies licensing years of their own operational records, such as support tickets, CRM histories, engineering reviews and project files, to AI labs and data buyers who need records of real work to train and evaluate AI agents.
Describe it with the discipline you would apply to any unclosed initiative.
- Describe stages, not proceeds. Report how many companies were screened, how many completed a data inventory, how many are in buyer review and how many signed. Do not forecast license values.
- Call it what it is. A license is typically a one-time payment for an exclusive AI training license for an agreed term. The company keeps ownership and nothing is sold.
- Keep it at the company level. Each company decides with its board whether to license. The fund does not own or sell portfolio data.
- Show the safeguards. A rights review comes before anything moves, de-identification and redaction requirements are agreed before any work begins, and data is delivered only after an executed agreement and the company's authorization.
- Disclose who gets paid. If anyone at the firm introduces a portfolio company through a referral program and could receive a reward, say so and explain how the fund documents treat it. The guide to management fee offsets and referral fees covers the LPA questions.
Timing questions follow. Licensing moves at the company's pace once its inventory starts, and how long data licensing takes within a hold sets realistic expectations.
A short script for the annual meeting
Illustrative wording; change every count and status to match what is true for your portfolio on the day you say it.
How to phrase it in LP reports
Small wording changes separate a credible report from one that invites follow-up questions you cannot answer.
| Instead of | Write (example phrasing) |
|---|---|
| AI will add significant revenue across the portfolio | Companies piloting AI in support and finance track cycle time and cost per transaction against a baseline |
| Our data is a major untapped asset | We are reviewing whether some companies can license historical records; no license is signed |
| Data monetization program | Data licensing review at company level, subject to each board |
| Expected proceeds of a stated amount | No proceeds projected until a license is signed and paid |
Keep all LP materials inside your fund's compliance review and marketing policies, which govern how initiatives and performance may be described. This is general information, not legal, tax or financial advice.
What it means if your firm is also a referral partner
Operating partners who screen the portfolio may introduce qualifying companies to SourceX as referral partners. Partners earn 25% of the eligible platform fees SourceX actually collects from the referred company's licensing deals, capped at $100,000 per referred company, paid only after the buyer pays and SourceX receives its fee. No reward is guaranteed, and it never reduces what the company receives.
Because LPs will ask, decide before registering whether the firm or an individual is the partner and how the fund documents treat any reward. The operating partner hub and the AI value creation playbook cover the wider role.
Limits and open questions
- AI value attribution is hard; LPs know it and discount answers that skip the baseline.
- Data licensing does not fit every company. Size, history, rights and an authorized sponsor all matter; the who qualifies baseline starts at 50+ full-time employees at peak (contractors excluded) and several years of documented operations.
- Proceeds are one-time and depend on buyers selecting the data, so they should never be presented as run-rate.
- Exposure reviews age quickly. An answer on which holdings AI threatens should carry a date and be refreshed before each fundraise.
Next step
Draft your answers in the four areas above before the next LP meeting, and add a staged line for any data licensing review. Map candidate companies with the network opportunity finder, and register as a partner if your firm decides to make introductions.
- Step 1Share your linkSend your personal link to a company you know.
- Step 2Company appliesThe company applies itself at /apply.
- Step 3Buyer selects and paysThe buyer selects and pays for the data and SourceX receives its fee.
- Step 4You get your rewardYour share of SourceX fees becomes payable.
Common questions
Should GPs include data licensing proceeds in fund projections?
No. A license is a one-time payment to a portfolio company, decided by that company's board, and it is payable only after a buyer selects the data and pays. Until a license is signed and paid, report the initiative by stage only. Including projected proceeds invites questions you cannot answer with evidence and weakens trust in the rest of the AI story.
How detailed should AI answers in a due diligence questionnaire be?
Specific enough to be checked: counts of portfolio companies by AI stage, the KPIs moved with their baselines, the governance policy and who owns it at each company, and dated examples. Attach the policy rather than describing it. Keep claims inside what your compliance team has reviewed, and update the answers before each fundraise.
What risks will LPs want covered if portfolio companies license data?
Expect questions on rights, privacy, confidentiality and reputation. A good answer explains that each company reviews its contracts and notices before anything moves, that de-identification and redaction rules are agreed before any work begins, that data is delivered only after an executed agreement and company authorization, and that each board decides independently whether to sign.
Who should answer AI questions at the annual meeting?
The operating partner or head of value creation should own the portfolio answer, ideally with one portfolio CEO describing a specific use case and its measured result. Deal partners take exit-readiness questions, and the CFO or chief compliance officer takes governance and reporting. Agree numbers and wording in advance so spoken answers match the written materials.
Does a referral reward earned by someone at the GP need to be disclosed to LPs?
Check the LPA first. Fund documents often address fees the GP or its affiliates receive in connection with portfolio companies, including whether they offset the management fee, and conflicts provisions may require disclosure or LPAC review. Ask fund counsel how a referral reward is treated before anyone registers, and disclose it consistently with the fund documents.
Related pages
- AI governance policy template, including who may approve licensing company data
- Management fee offsets and referral fees: what sponsors need to check in the LPA
- How long does data licensing take within a PE hold period?
- Referral opportunities for private equity operating partners
- AI value creation in private equity: a playbook for operating partners
- Which US businesses are a fit for a SourceX data licensing introduction
Free resources
- MCP ROI calculator — Estimate hours saved, implied savings and first-year ROI from MCP.
- Business exit readiness assessment — A preliminary exit readiness score and checklist for advisors.
- SDE vs EBITDA calculator — Seller's discretionary earnings next to market-rate EBITDA.
- All free tools · MCP resource center
By SourceX Partnerships Team · Published 2026-10-09 · Updated 2026-10-09
Know a US company with valuable proprietary data?
Become a referral partner from anywhere we support, get your link and introduce an owner or authorized decision-maker.
Refer a company →I own a business
Explore licensing your company's data to AI developers worldwide. Start a short assessment; no uploads needed.
Start an assessment