Researching Funders & Grant Opportunities with AI

Writing a brilliant proposal to the wrong funder is wasted effort. The most successful grant writers spend serious time up front finding funders whose priorities genuinely match their work — because a good match is the single biggest predictor of getting funded. AI makes this research dramatically faster, but it also introduces a real risk: AI loves to invent funders that do not exist. This lesson teaches you to research efficiently and verify ruthlessly.

What You'll Learn

• How to use AI to build a shortlist of likely funders
• How to assess whether a funder is a genuine fit
• The critical verification step that protects you from AI hallucinations
• A prompt workflow for scanning a funder's website in seconds

Why Funder Fit Matters More Than Anything

Foundations and grant programs each have a "personality" — causes they care about, geographic areas they serve, the size of grants they give, and the type of organizations they support. A youth literacy foundation will not fund your river cleanup, no matter how beautifully you write. The skill of finding funders whose priorities align with your mission is called prospect research, and it is where AI delivers an enormous head start.

Your goal is a shortlist of 10–20 realistic prospects, each scored for fit, so you spend your writing time only where it can pay off.

Step 1: Build a Shortlist With Perplexity

Because funder research demands verifiable facts, Perplexity is the right tool — it cites sources you can click. Try a prompt like this:

List 12 private foundations that have awarded grants between $10,000 and $75,000 for {youth mental health programs in the Pacific Northwest} in the last three years. For each, give: foundation name, focus areas, geographic scope, typical grant size, whether they accept unsolicited proposals or letters of inquiry, and a link to their grants page.

Perplexity returns a list with links. Treat it as a list of leads to investigate, not confirmed facts. Some entries will be accurate, some outdated, and some possibly invented. That is normal and expected — verification is the next step, not an afterthought.

Step 2: Verify Every Funder (Non-Negotiable)

This is the step beginners skip and regret. AI confidently fabricates foundation names, giving histories, and contact details. Before you invest a minute in writing, confirm each prospect against a primary source:

• The funder's own website — the single best source for current priorities and deadlines
• Candid's Foundation Directory — the gold-standard nonprofit funder database
• The funder's IRS Form 990 (for U.S. foundations) — public tax filings that show exactly who they have funded and how much
• GuideStar / ProPublica Nonprofit Explorer — free places to look up 990s

If you cannot find a funder through any of these, assume the AI invented it. Never include an unverified funder in your plan, and never reference a "past grant" the AI claims a funder made unless you have confirmed it independently.

Step 3: Scan a Funder's Website Fast

Once you have a verified prospect, you still need to understand what they want. Copy text from their grants or guidelines page and hand it to Claude or ChatGPT:

Below is text from a foundation's grants page. Tell me: (1) what they fund and what they explicitly do NOT fund, (2) their typical grant size, (3) geographic restrictions, (4) deadlines and whether they want an LOI first, (5) any eligibility requirements, and (6) clues about their values or language I should mirror in a proposal. Page text: {paste}.

In seconds you get a clean briefing that would have taken twenty minutes to assemble by hand — including the often-overlooked "what they do NOT fund" list, which saves you from disqualifying mistakes.

Step 4: Score Each Prospect for Fit

Turn your shortlist into a ranked plan. Paste your verified prospects and your program details:

Here is a list of verified funders {paste} and a summary of my program {paste}. For each funder, rate the fit from 1 to 5 and explain why in one sentence. Then rank them and tell me which three I should pursue first and why.

Now you have a prioritized action plan instead of an overwhelming list. You will write your strongest proposals first, to the funders most likely to say yes.

A Realistic Example

A community arts nonprofit wanted local foundation support. They used Perplexity to surface 14 possible funders in their state. Verifying against Candid and 990 filings, they found that 4 were invented or no longer active, 3 had shifted away from arts funding, and 7 were genuine matches. AI turned a multi-day research slog into an afternoon — but only because they verified. Had they trusted the raw list, they would have wasted weeks chasing foundations that did not exist.

Pitfalls to Avoid

• Phantom funders. Always verify existence and current priorities before writing.
• Stale information. AI training data lags reality; deadlines and focus areas change. Trust the funder's own current website over the AI.
• Ignoring the exclusions. A funder's "we do not fund" list matters as much as what they do fund.
• Over-broad searches. Narrow by cause, geography, and grant size for sharper results.

Key Takeaways

• Funder fit is the biggest predictor of success — invest in prospect research before writing
• Use Perplexity to build a shortlist because it cites sources you can verify
• Verification is non-negotiable: confirm every funder against their website, Candid, or IRS Form 990 filings
• Use Claude or ChatGPT to summarize a funder's guidelines, including what they will NOT fund
• Score and rank prospects so you write your strongest proposals to the best-fit funders first