Research Sources — Where Funder Data Lives

Funder research is detective work. The information you need is scattered across multiple sources, and no single database has it all. Knowing where to look — and what each source is good for — is a foundational skill.

IRS Form 990-PF: A Solid Starting Point

Every private foundation in the U.S. files Form 990-PF with the IRS. These are public documents, and they’re one of the best places to start understanding what a funder actually does.

What you can learn from a 990:

Who they funded — specific grantees, amounts, and purposes
Total giving — how much the foundation distributed that year
Assets and endowment — how much they have to give
Officers and directors — who makes decisions
Trends — compare multiple years to see if a program area is growing or shrinking

Pro tip

Where to find 990s: ProPublica’s Nonprofit Explorer, Candid/Foundation Directory Online, and the IRS Tax Exempt Organization Search all provide free or low-cost access.

The limitation of 990s is that they’re filed on a lag — you’re often looking at data that’s 1-2 years old. And the “purpose of grant” descriptions are sometimes vague. 990s are a strong starting point, but they’re not the full picture. You’ll need to combine them with other sources to build real confidence in a prospect.

Traditional Grant Databases

Several subscription and free databases aggregate funder information. These are the tools most grant professionals have historically relied on:

Grants.gov

The central portal for federal grant opportunities. Free. Essential if you pursue government funding.

Candid (Foundation Directory Online)

The most comprehensive traditional foundation database. Subscription-based, but many libraries provide free access.

GrantStation

Curated database of foundation, corporate, and government grants. Subscription-based.

State-specific portals

Many states maintain their own grant databases for state-funded opportunities.

Watch out

These databases are valuable, but the workflow is manual and time-intensive. You’re searching by keywords and filters, scrolling through pages of results, reading individual funder profiles, and copying relevant details into a spreadsheet — one funder at a time. A thorough search through a single database can take hours, and you’ll need to repeat the process across multiple sources to get a complete picture. It’s effective work. It’s also slow work.

Funder Websites

A funder’s own website is often the most current source for:

Current funding priorities and program areas
Application guidelines and deadlines
Recently funded projects and grantee profiles
Strategic plans that signal where they’re headed

The guidelines tell you what a funder says they want. The list of recent grants tells you what they actually fund. When those two don’t match, trust the grant list.

Your Network

Don’t underestimate word of mouth. Some of the best funder intelligence comes from:

Peer organizations that have been funded by a foundation you’re considering
Professional associations and grant writing communities
Conference sessions where funders speak about their priorities
Board members and volunteers who have connections in philanthropy

Example

A colleague saying “I talked to their program officer last month and they’re really interested in workforce development right now” is intelligence you won’t find in any database.

AI-Native Research: A Different Approach

Traditional databases were built for humans to search — you type keywords, apply filters, and read through results. That works, but it means you’re limited by the search terms you think of and the time you have to read.

A newer approach is databases built for AI to search. Instead of you querying by keywords, AI compares your organization’s full profile against funder records — matching on mission, geography, program type, funding patterns, and giving history simultaneously. The AI reads and cross-references at a scale no human can, and it surfaces connections you might never find through keyword searches alone.

In Grantable

Grantable’s GrantGraph is built this way — a dataset designed from the ground up for AI-powered research rather than manual browsing. It draws from 990 filings, public grant records, and organizational data, and it’s structured so AI can evaluate fit across multiple dimensions at once. You describe what you’re looking for, and the AI searches the way a researcher would — but across hundreds of thousands of funders instead of a few dozen.

The result isn’t that traditional research becomes unnecessary — your network, funder websites, and direct conversations are still essential. But the hours spent in manual database searches compress significantly, freeing you to spend more time on the sources that require human judgment.

Cross-referencing sources for confidence

When AI surfaces a potential funder, the strongest validation comes from triangulating across sources: the 990 confirms actual giving history, the website confirms current priorities, and your network confirms relationship potential. Any one source can be outdated or incomplete — the combination is what builds confidence in a prospect.

Check your understanding

You find a foundation whose website says they fund 'education innovation,' but their last three 990s show grants exclusively to after-school tutoring programs. What should you conclude?

Key Takeaways

990 filings are a strong starting point for understanding what foundations actually fund — but not the full picture on their own
Traditional databases are valuable but manual and time-intensive — keyword searching, reading, and spreadsheet-building one funder at a time
Your professional network provides real-time intelligence that no database can match
AI-native research tools search across your full organizational profile, compressing hours of manual searching into minutes

Next Lesson

You know where the data lives. Next, let’s look at how AI funder matching actually works — what it’s good at, and what you still need to check yourself.