∞Direct answer
Direct answer
An AI match is a lead to test. It is not investment advice. It does not say a deal will close or pay out. Before a meeting, both sides should check key terms. Mark data as observed, missing, or inferred. Ask why the match was made and what it may miss.
Check one match, not your whole list
Use this guide when an AI platform suggests that one founder and one investor should meet. It is not for making a long list or sending outreach. Ask one thing: is this talk worth both sides’ time?
AI can sort data. It cannot show that a deal will close, that an investor will lead, or that an investment will pay out. NIST says AI risk work needs care in design, use, and tests. Do that here. Read the result and make your own call. [1]
This short check is not full due diligence on an investor. Due diligence means checks before an investment. It is not regulated investment advice.
Check the key terms first
Start with both sides. Check stage, normal cheque range, country and legal scope, sector or model, lead or follow role, fundraising date, and time to meet. Mark each point: confirmed, unconfirmed, or not a fit. A good profile does not prove these points.
Then check for a portfolio clash and who can say yes. A founder can ask if the firm backs a direct rival, how it will handle data, and who can move a deal on. An investor can ask if the founder needs a lead, expert help, customer links, or market views. The British Business Bank also tells founders to check fit and conflicts of interest. [2]
Label observed, missing, and inferred data
Use three tags. Observed is a new fact with a source. An investor may list a sector. A founder may name a round. Missing means the system has no data, such as fund cash left or a private rival deal. Inferred means the AI made a view from clues. Past work in a field may hint at useful help.
Do not show inferred data as fact. Ask when data was checked, who gave it, and if either side can fix it. If a key point is missing, send a clear note first. The meeting can wait. This record is a guide, not a tested score model.
Ask for clear reasons, not one score
A good match note gives the few reasons for the match. It also says what may make it weak. For example: stage looks right; country is confirmed; lead interest is not known; a portfolio clash needs to be shared. Keep a profile claim apart from a fact said in a call.
Do not rely on one percent or call a match ideal. Fit can shift with goals, fund pace, who can decide, and rapport. Ask what data was not used, what was inferred, and what would change the view. A clear gap is better than fake detail.
Do a 15-minute pre-meeting review
1. Compare key terms. 2. Read the data tags. 3. Find one clash or gap to clear. 4. Write the meeting aim in two lines. 5. Choose: introduce now, ask first, or decline with care. Write down why. This can help later matches without overstating the data.
Example: a climate-software founder seeks a pre-seed round. An investor page says climate and early-stage investment. The match has gaps. The cheque range and lead role are not known. A firm it backs may sell to the same buyer. Ask about the clash and scope. Do not assume an introduction.
Use the first meeting to test fit
If the checks back a meeting, send a short note. State the round, date, aim, and one point that could change fit. A founder does not need a full data room for this first talk. An investor need not say yes early. Do due diligence later. It should fit the stage and the link between the two sides.
After a term sheet, more files may matter. YC lists firm records, finance, IP, and key deals. Keep that later work apart from an early AI match. [3]
FAQ
Frequently asked questions
Should an AI match decide if I take a meeting?
No. Use it to set up the choice. Then check your terms, conflicts, and judgment. A match is not investment advice. It does not predict funding or returns.
What missing field should I check first?
Check the gap that could make the meeting pointless. It may be stage, cheque range, country, lead role, a portfolio clash, or timing. It depends on the case.
How can a founder ask about a portfolio conflict?
Ask in a short, direct way. Do not ask for confidential facts: “Do you have an investment that could conflict with this deal? How do you handle it?”
Can an investor use an inference to skip a meeting?
An inference can prompt a question. It cannot replace a check. State the concern and check facts when you can. Do not turn an inference into a claim about the founder or firm.
REF
Sources
- 1
National Institute of Standards and Technology
AI Risk Management Framework ↗ - 2
British Business Bank
How to conduct due diligence on investors ↗ - 3
Y Combinator
Series A diligence checklist ↗