Why your nonprofit doesn’t show up in ChatGPT or Google’s AI answers
Most of the organisations I work with have already done the obvious things: a fast, accessible, well-ranking website. So when a board member asks ChatGPT or Google’s AI Overview about their sector and the organisation doesn’t appear, it feels like a contradiction. It isn’t. Being findable and being citable are two different jobs.
Search rankings and AI answers are graded differently
A traditional search result rewards a page that matches a query and earns enough authority to outrank the alternatives. An AI answer engine is doing something else. It is assembling a response from a small number of sources it judges trustworthy enough to summarise or quote directly, and it strongly prefers sources it can parse with confidence: clear authorship, structured data that states plainly what a page is and who wrote it, and content that reads as a direct, well-supported answer rather than a persuasive pitch.
A page can rank respectably in Google and still never be selected as a source for a generative answer, because ranking and citation draw on overlapping but distinct signals. For most nonprofits, the gap is not the writing. It is the plumbing underneath it.
What AI answer engines are actually looking for
Generative engine optimisation, GEO, is the practice of making a site legible to these systems, as distinct from search engine optimisation, which targets rankings. The two overlap, but GEO leans harder on three things: structured data that states plainly what a page is, clear and verifiable authorship, and corroboration from sources outside your own site.
The three gaps I see most often
No structured data on the content that should be most citable
Blog posts and guides are exactly the content an AI system wants to summarise, but if they carry no Article or FAQPage schema, there is nothing telling the system what the page is, who wrote it, or when it was last checked. It has to guess, and it usually guesses in favour of a competitor’s page that made the answer easier.
No llms.txt file
This is a newer, still-emerging standard, but it is a low-cost signal: a plain-text file that tells AI crawlers what your site is about and points them to your most authoritative pages. Almost no nonprofit sites have one yet, which makes it one of the cheaper wins available.
Thin third-party presence
This is the one that matters most and costs the most to fix. AI systems weight sources that are corroborated elsewhere: directory listings, sector citations, other organisations linking to you, coverage in the outlets your sector actually reads. A single well-built page rarely earns a citation on its own merits. It earns one because other sources vouch for it.
What to fix first
- Add Article and FAQPage schema to your highest-value content first, particularly anything that already answers a specific question well
- Add a llms.txt file. It takes an afternoon and there is no reason not to
- Treat backlinks, directory listings and sector citations as the main event, not an afterthought
None of this replaces third-party authority. Structural fixes get a page into consideration for citation. Backlinks and sector coverage are still what earns it.
Questions your board should be asking
- Do we know whether ChatGPT, Gemini or Google’s AI Overviews mention us at all, for the queries our funders and clients actually ask?
- Does our content carry structured data, or is it just formatted for human readers?
- Who outside our own website vouches for us, and would an AI system be able to find that?
- Is this framed as a marketing project, or as part of how we protect our visibility to the people who need to find us?
Not sure where you actually stand?
A Digital Capacity Diagnosis covers search visibility alongside accessibility and governance, including whether your content is structured in a way AI systems can use. $1,500 + GST.
Book a Diagnosis →
