We Raised $101K in 3 Weeks: Here’s Our AI-Assisted Board Fundraising Playbook

We ground out 58 grant applications over one year to net $108,000. Then we amazed ourselves by raising another $101,000 from individuals in just three weeks.

In March 2026, I sat with my wife Sally at our dining room table in Oakland with the donation list open between us, refreshing the total more often than I’d like to admit. We counted up the checks from our three-week individual donor campaign. And then we counted them again, because it didn’t seem quite real:

• 73 supporters
• $101,000
• Median gift: $100
• Top gift: $25,000
• Almost none of it restricted

The total was surprising, but so was the gift pattern. It was a lot of people giving what they could and one or two people stretching farther than we expected. Most surprising, the board fundraising campaign resulted in almost the same amount of money we’d raised in a whole year of institutional grant applications:

• 58 grant applications
• 7 yeses
• $108,000

Sally and I co-founded Outdoorithm Collective in 2024 to take families who’d never been camping into the outdoors. It’s a startup nonprofit, not yet two years old, run by just the two of us. It’s a hard mission to fund, even though it’s an easy one to feel. Most people intuitively understand the impact of a first camping trip, picturing a city kid seeing a night sky full of stars for the first time. The 58 applications we ground out in 2025 mattered, and the seven yeses funded a beautiful season.

Although I spent a decade in tech philanthropy at Google.org, that experience failed to teach me a lesson the board campaign in the spring taught me quickly: Grants move when institutions are ready. Individual gifts move when relationships are ready. If Sally and I want a strong development shop, we need to run both as deliberate systems. Most nonprofits run grants like a program but their own network like an afterthought.

That’s the gap this post is about. If your organization has a board of directors or even a small group of champions, you probably have more relationship capital than you’ve had time to map.

The Overlooked Asset: Your Network 

If you don’t yet have a system to connect with the combined LinkedIn and personal networks of your founder, executive director, board, and senior staff, now’s the time.

I’ve sat on enough boards to know what board members say when they sign on: “If you ever see anyone in my LinkedIn who can help, just tell me. I’m happy to make introductions.” They mean it. But it almost never produces an introduction.

The reason isn’t that board members don’t want to help. It’s that saying “I have 7,000 LinkedIn connections” is useless to a development director. Nobody has time to scrub 7,000 names against a mission. So, the board member waits to be asked. The development director never asks. The asset sits.

This is where AI became really helpful for us. Not to replace relationships, but to surface the relationships we were already sitting on.

Your database does part of this. It tells you which of your current donors have capacity you haven’t tapped yet. The other part is the network play. It finds the people your board and your team already know but have never been in your database.

The Play, in 4 Steps

Here is the workflow Sally and I ran for Outdoorithm Collective. We know we didn’t invent some kind of fundraising revolution here. The approach we’re sharing blends the best of peer-to-peer fundraising with the kind of prospect analysis a data team would run at a larger organization. But because our lean two-person operation has more ideas than hours, we needed AI to close the gap. We’ve since run versions of this method for several other nonprofits, and the shape holds.

1. Export the networks.

Every board member can pull their full LinkedIn connection file in a few clicks, under “Get a copy of your data” inside LinkedIn settings. What comes back is a CSV with names, companies, profile URLs, and sometimes emails, for connections who have made theirs available. Treat the file like donor data once it lands, because that is what it has become. Decide who on your team gets access, where it lives, and how long you keep it. For Outdoorithm Collective, we started with just Sally and me, the two founders. That alone came to about 4,000 deduplicated connections.

2. Enrich every name.

For each connection, pull publicly available signals. Honestly, it’s easy to get carried away in this step, so we were disciplined, looking only for signals that were truly useful for our individual campaign. Career history. Board seats. Documented philanthropy. Public political giving. Published writing or interviews, particularly on topics related to outdoor access or urban youth. Estimated giving capacity from public records. Most of this comes from public or permissioned sources, but public is not the same as casual. Use signals you can stand behind, document where they came from, and keep humans in control of how the data is read.

AI helped with the first pass. It was a much faster process than we could have done manually, but this step absolutely needs human review, especially when signals are easy to misinterpret. How do you know if a signal is a little thin? For us, it was helpful to notice a contact who once worked in education or who gave publicly to an environmental cause. But was that enough for us to assume they cared about our mission in particular? No, not on its own.

3. Score three things.

For each enriched person, score three dimensions:

• Capacity: What can this person plausibly give at the high end of their generosity?
• Affinity: Does anything in their public life suggest they care about what we do?
• Relationship: How close is this person to someone on our team? An acquaintance, a real friend, or somewhere in between?

The first two are jobs AI can do well from public data. The third is harder. There are two ways to do it, and both deserve real attention and care.

If you connect your email, calendar, and text messages, AI can read how often and how recently you actually have talked to someone, and use that as a measure of closeness. We did that for Outdoorithm Collective.

If you would rather not connect your messaging and meeting systems to protect privacy, you can grade closeness by hand. Your team rates how well it knows each strong match, one row at a time. We did that too. It was the slowest step, and the one that made the rest of it work.

With either method (or, in our case, using both), Sally and I had to sit with the list and each board member to ask the question AI couldn’t answer: Are we actually comfortable reaching out to this person for a donation? Once we completed our human review to decide who counts as a real relationship, we narrowed our list to about 300 people. These were not just promising names in a spreadsheet. They were real relationships we could responsibly act on.

4. Build the bridges.

For every high-affinity, high-capacity person with whom someone on our team had a real relationship, we sent a personal ask. The system gave us a one-paragraph briefing on each one: their career, what we knew about their philanthropy, the natural opening. We wrote the actual asks ourselves. Some were a text message. Some were a call. A few were conversations over a cup of coffee, when the ask came only after we’d chatted about kids, work, and why our mission matters so much. We delivered about 300 asks in three weeks.

I shared the results in the opening of this post: 73 gifts. $101,000. About one in four people we asked said yes. We’re not claiming every organization following our playbook should expect a 25% conversion rate, but it confirmed how much more potential exists in a truly warm relationship versus a cold prospect.

Set a Few Guardrails Before You Start

Treat the network export and the enrichment data like donor records, because that’s what they’ve become. Limit access. Decide where those records live and how long you’ll keep them. When the team has finished reviewing, move the reviewed list, the connection notes, and the next steps into your CRM. The point is to turn hidden relationships into trackable opportunities, not to run a shadow fundraising system in a spreadsheet on someone’s laptop.

And keep AI’s job narrow. Its task is to help your team notice where a real relationship may already exist. That’s it. Don’t automate a personal ask. Don’t let AI decide who matters. Don’t confuse a high capacity score with generosity, or a strong public signal with a real friendship. The human on your team—the one who actually knows the person—makes the call. Every time.

What This Looks Like with a Board Chair

A few weeks after our campaign closed, I sat down with the board chair of a Bay Area education nonprofit. We ran the same play on his LinkedIn network in one sitting. Several thousand connections became 100 mission-aligned possibilities. He went through that list with me, top to bottom. By the end, he had circled 10 connections he was personally willing to approach. This is the turning point, when your board member shifts from saying, “Maybe this person” and starts writing a list of people to email the same day.

That’s the version of board engagement most development directors describe when you ask them what they wish for. Not a polite “I’ll think about it.” A board member taking responsibility for a specific list of specific people, with a concrete next step the staff can actually run.

3 Things to Do This Week

If you’re reading this between meetings with a campaign deadline looming, you don’t need to start with a board list of 4,000 names. Start with one person if that’s all you can do. If you can be slightly more ambitious, follow this scaled-down version of the play, giving yourself a deadline of just a few weeks.

1. Pick one trusted person.

Founder, ED, board chair, longtime senior advisor. Someone with a real network and skin in the game. Start there. Don’t try to do the whole board at once.

2. Ask for the LinkedIn data file.

A few clicks under “Get a copy of your data” in LinkedIn settings. Treat the file like donor data the minute it lands.

3. Run a pilot, not a program.

Here’s how to start running your pilot: Pick one name from your board chair’s LinkedIn export. Say it’s Sarah, a partner at a Bay Area law firm.

Open Claude (or ChatGPT, Perplexity, or Gemini) and ask for a paragraph of public context: Sarah’s career, board service, public philanthropy, public writing. Tell it to cite every claim with a source URL or respond with UNKNOWN. That gets you maybe 70% of what you need.

Three free public sources fill in the rest:

• ProPublica’s Nonprofit Explorer searches IRS Form 990s by name. If Sarah shows up as a trustee or officer at a foundation, she’s already giving her time to a mission. This is the strongest affinity signal you can get for free.
• FEC.gov lists individual political contributions; $5,000 across recent cycles tells you she writes checks at that size when something matters to her.
• Zillow, plus a people-search like 411.com to find Sarah’s address, gets you a home value—and roughly 5% of estimated home equity is a defensible soft floor for a first major ask.

Plan for 10 minutes per row, once you have the rhythm. The next 24 go faster.

If you’d rather not do this by hand, we built Kindora to run every one of those lookups automatically and assemble the score in one pass. The point isn’t which tool you pick. It’s that you pick one and set it up to repeat. Inside Claude, the free Kindora connector also lets you search 174,000+ foundations and their giving patterns.

Whatever path you choose, four rules keep AI honest in this work:

• Make it cite a source for every claim or respond with UNKNOWN.
• Don’t accept a name match alone—common names need an employer, city, or board overlap before you trust them.
• Downgrade your confidence when a high estimate is sitting on a thin signal; “Could give $100K” backed only by a job title is a guess, not an estimate.
• And when you draft the ask itself, pull only from your real story bank, not from anything AI invented to sound persuasive.

Grants Still Matter

As effective as it is, AI-assisted fundraising from a board member’s personal network is not a replacement for institutional giving. Outdoorithm Collective is still applying for grants this year. So is that education nonprofit. Grants still matter, and they can fund parts of the work nothing else will.

The same network helps you there too. It shows you who can open a door at a foundation or a corporate funder, which is a warm path into institutional money.

What this adds is the other engine. The one that runs on relationships you already have, with tools that finally exist.

For 10 years at Google.org I helped direct close to $700 million into nonprofit work. The part of that decade that aged the best in my memory was almost never the institutional grant. It was the human moment where somebody close to the work asked somebody close to them, and that person said yes.

As a founder, I understand that AI doesn’t create that kind of generosity. It simply makes generous supporters more findable. An AI assist plus the steps in our playbook helped us become more attentive to the generosity already within reach. That attention turned three weeks of work into 73 gifts.

What could your board’s network make possible for your mission?