AI Prospect Research: What It Can (and Can't) Tell You About a Donor

Prospect research has always been part art, part legwork — piecing together public information to get a sense of whether someone has the capacity and inclination to make a meaningful gift. AI has made the legwork side dramatically faster. It hasn't changed what the information actually means, and it's worth being honest about where the line sits.

What it's genuinely good at

Speed and volume. Researching capacity indicators on fifty donors used to mean fifty separate manual searches. AI lets me process that volume in a fraction of the time, which means a capacity review that used to take a week can often happen in a day.

Pattern recognition across a data set. AI is very good at noticing things a person might miss when they're looking at one record at a time — a subtle uptick in giving that suggests growing capacity, or a cluster of donors connected to the same event or program that hints at an affinity group worth cultivating together.

Organizing scattered public information. Capacity indicators are often spread across a dozen different sources. AI is useful for pulling that together into a single, organized starting point, rather than a researcher manually cross-referencing tabs for an hour per donor.

What it can't tell you

Whether someone actually cares about your mission. Capacity is only half the equation. AI can tell you someone likely has money. It cannot tell you whether they have any emotional connection to your organization specifically — that still comes from relationships, notes, conversations, and institutional memory that no dataset captures.

The right way to approach someone. Every major donor conversation is different, and the right opening, tone, and ask depends on the relationship and the moment, not a data profile. This is a judgment call, every time, and it's the part of the work I'd never hand off.

Context that isn't public. A donor's public profile might look impressive while their actual current circumstances — a recent business downturn, a family situation, competing philanthropic commitments — aren't visible anywhere online. Good research treats AI output as a starting hypothesis, not a conclusion.

How I actually use it

I treat AI-assisted prospect research the same way a sales team treats lead scoring: a fast, useful first pass that tells me where to spend my limited time, not a replacement for actually picking up the phone or sitting down with someone. The research narrows the list. The relationship — and every ounce of judgment about how and when to make an ask — is entirely human, every time.

If a consultant or a tool is telling you AI can tell you exactly who to ask and exactly what to say, be skeptical. The best it can do is get you to the right conversation faster. The conversation itself is still the whole job.