Every small business is in the fundraising business, even when nobody calls it that. The sponsorship pitch to the local league. The customer appeal after a rough month. The launch email. The community drive, the school partnership, the end-of-year thank-you to your best customers. Constant, small, emotionally loaded asks — and most owners either avoid them entirely or burn an afternoon writing each one from a blank page.
There's a clean division of labor worth borrowing for all of it. GoFundMe CEO Tim Cadogan, in a recent MIT Sloan Management Review conversation, described how the platform uses AI: it draws on lessons from tens of millions of fundraisers to help people tell their stories and mobilize communities — while the people making the asks keep the emotional core. His summary was five words: "AI for process, people for heart."
That's not a slogan. It's an implementation plan, and a small business can run it in an afternoon. Here's the translation.
The Two Layers, Defined
Take any ask your business makes — a sponsorship request, a launch announcement, a community campaign — and it splits into two layers:
The process layer is everything structural: the shape of the ask, the draft, the sequence of follow-ups, the logistics of who gets contacted when, the receipts and thank-yous, and the comparison of what worked versus what didn't. This layer rewards scale and pattern-matching, which is exactly what AI is good at.
The heart layer is everything specific: the one detail that only you know, the reason this particular ask matters to this particular community, the voice your regulars would recognize. This layer punishes averages. An AI can imitate warmth, but it cannot know which customer waited out your renovation or which sponsor has quietly covered your league's jerseys for three years.
The mistake most businesses make isn't using AI too much or too little — it's assigning the wrong layer. AI gets asked to supply the heart (and produces polished, interchangeable sentiment), while the human gets stuck doing the process work (manually tracking who was contacted, when, and what happened).
Flip it.
How to Run the Split, Step by Step
1. Have AI build the story spine
Every effective ask has the same skeleton: the situation, the need, the stakes, the specific request, and the deadline. Ask your AI assistant to outline the spine for the type of ask you're making — a sponsorship pitch, a launch email, a community appeal. The AI isn't writing your ask yet; it's giving you the load-bearing structure so you never start from blank.
2. Feed it facts, not feelings
Give the AI the concrete inputs: who the ask is for, what you're requesting, what they get, dates, numbers, and any past history between you. Then have it draft from proven ask structures — direct ask, story-first, offer-first. The more factual the inputs, the less the draft has to fabricate, and the less fabricated warmth you'll have to cut later.
3. Your edit pass is the heart pass
Edit the draft with one question: could any business in my town have sent this? If yes, it has no heart yet. Add the detail only you have — the name of the specific field the sponsorship banner would face, the customer who asked when you'd reopen, the reason this drive matters this year and not last year. Two or three specifics do more emotional work than any amount of adjectives.
4. Automate the process end-to-end
This is the layer businesses leave manual, and it's where the compounding lives. Once the ask is sent, the process layer should run itself: follow-up reminders for non-responders at day 5 and day 12, receipts and confirmations, thank-you notes triggered on a yes, a scheduled recap after the campaign closes. Ask your AI assistant to set up the sequence once; after that, the ask works its list while you work the business.
5. Close the loop on what worked
GoFundMe's version learns from tens of millions of fundraisers. Your version starts smaller: after each ask, note what happened — response rate, yes rate, what the actual yeses said — and feed it back. Three or four asks in, you have a structure library: "the story-first version out-pulled the direct version two-to-one with this audience." That library is an asset. It's the thing competitors can't copy, because it's built from your community's actual behavior.
Where the Split Breaks
Two honest limits, so you plan for them rather than get surprised:
AI drafts toward the average. Patterns learned from millions of asks describe what works for most people, most of the time. Your community is not an average — it's a specific group with a specific history with you. That's precisely why the heart pass exists, and why it can't be skipped on the asks that matter most.
Mind what you feed it. If an ask involves customers' personal situations — someone's illness, a family's hardship, a staff member's story — get their okay before their details go into any AI tool, and check what the tool does with your data. The process layer should never create a privacy problem on the way to saving you an afternoon.
The Monday Version
- [ ] List the next three asks your business needs to make (sponsorship, launch, appeal — anything)
- [ ] Have your AI assistant outline the story spine for each and draft from your facts
- [ ] Do the heart pass: add two specifics per ask that only your business knows
- [ ] Set up the follow-up sequence once — day 5 nudge, day 12 nudge, thank-you on yes, recap at close
- [ ] After each ask closes, record what happened in one line, and feed it back next time
The point of the split isn't to make your asks faster, though they will be. It's that the asks stop being the thing you avoid — because the machine handles the machinery, and the only part left for you is the part only you can do.
The Small Business AI Starter Kit includes step-by-step setup guides for exactly this kind of process-layer automation — drafting, follow-up sequences, and tracking — built for small businesses.
Top comments (0)