Forty-four hours into our annual epic travel session to Indonesia, and we finally had a breakthrough on our giveready.org B2A project. We left our Swiss apartment on Friday evening, and we won't touch down at our summer digs until midday Tuesday, that's why I call it epic travel. You get in the travel zone, delays don't bother you, walking around the cabin of the plane for a 13-hour flight is part of the deal. Suck it up, get used to it. Commit.
Midway through, in an airport hotel 8 hours before flying on to the next leg, Rote Island, Nusa Tenggara, Indonesia. I open the daily digest the way I have every morning for the last forty-two days and read the same small table. Ten questions, ten answers, and a column that has said n nearly all the way down since the end of May. This morning one row said y, which means it wasn't 0: we had our first citation from an agent.
An AI search engine cited our charity directory as a source. It happened once before, for a single day at the end of June, then vanished when the index turned over. This one I want to understand, because it is the first piece of hard evidence for what answer engine optimization can do for a small nonprofit, and it did not come from where I expected.
What does this all mean? For anyone landing here cold, SEO gets you ranked on Google. AEO and GEO, answer engine optimization and generative engine optimization, are about being the source the AI assistants quote when someone asks them a question. A growing share of donors no longer search "best youth charities" and click through ten blue links. They ask ChatGPT or Perplexity or Claude and read one answer. If the engines never cite you, you are not part of that AI conversation at all.
Since late May I have been running a fixed set of ten donor questions through an AI search engine every morning and logging whether GiveReady, the nonprofit directory I have been building in public since April, gets cited. The questions never change. That is deliberate: if I rewrote them until we won, the win would be worthless. For forty-two mornings the score was zero out of ten. Not cited-but-losing. Absent. The engines built their answers from the big rating sites, established charity listicles and the charities' own websites, and we were nowhere in any of it. I retired the first theory behind this project back on Day 63, and this stretch since has been the long unglamorous kind I wrote about in building in public is hard.
The citation did not come from the question I expected. Not the big American youth-charity question, where the established raters own the answer. Not music education, where a national nonprofit and a well-known listicle site sit on every response. The question that broke was about charities supporting township youth in Cape Town. When I looked at who we beat for that citation, it was Instagram pages, Facebook, Reddit threads and a city tourism magazine. Nobody credible had ever answered that question properly. So when a structured guide built from verified data existed, the engine took it.
There are tens of thousands of registered charities, and the rating industry covers a thin slice at the top. Everything below that, the local surf therapy programme, the township art project, the small adventure fund, is invisible to the AI engines because nobody has written anything worth quoting about them. On the big categories we lose to twenty years of accumulated authority. On the niche causes there is no incumbent to lose to. The answer space is empty, and empty ground is where a two-person operation can actually win.
That's also where the charity grant application we wrote in April finally gets a nod. The whole pitch was that AI can help increase donations for the smaller charities, and this is the first real sign of it.
I am not going to publish the full method here. Partly because this is still one citation on one morning and I would look silly claiming a system off a sample of one. Partly because the process itself is turning into know-how worth keeping. The direction, though, is public by definition. You are reading it, and we are sharing it so ideally others can learn in the process as well about how to promote their small charity.
One more reason this matters to me personally. The Finn Wardman World Explorer Fund, the fund we set up in our son's memory to get young people out into the world, now has full 501(c)(3) status. Adventure and youth empowerment are exactly the kind of niche causes the rating sites never touch. The next guides get built there.
The guides so far, if you want to see what structured charity data looks like:
- Best surf therapy charities for at-risk young people
- Best UK youth charities for the outdoors and at-risk teenagers
- Best US youth charities for mentorship, sports and skills
- Charities supporting township youth in Cape Town
- Best music education charities for underprivileged kids
- Best youth mentoring charities to donate to, US and UK
If you run a small nonprofit, try asking an AI assistant about your own cause sometime. If the answer comes back as Instagram links and a Reddit thread, that is not a locked door. That is empty ground nobody has claimed yet.
Tomorrow at eight the digest gets read again. The citation may hold or it may vanish, the engines rebuild their indexes all the time. Either way the table finally has a y in it, and I know where to dig next.
Monthly Revenues $11,000 | Clients 2 | Prospects 1 (AI outbound agent now live)