AI Visibility Is the New SEO for Indie Makers
Search has quietly changed.
People aren’t clicking links as much anymore. They’re getting answers directly from AI tools like ChatGPT, Google AI Overviews, and Perplexity.
If your product isn’t mentioned in those answers, you’re invisible.
What’s actually happening
More users start searches in AI tools
A majority of searches now end with zero clicks
AI summaries are replacing “research via links”
For indie makers, this means ranking on Google isn’t enough anymore.
SEO → GEO (Generative Engine Optimization)
The goal isn’t to rank pages.
The goal is to be included in AI answers.
Ask yourself:
“If someone asks AI for the best tool for my problem, would it mention me?”
If not, that’s the new gap.
What helps AI pick you up
Clear, structured content (short paragraphs, real FAQs)
Specific use cases and comparisons (not generic marketing)
Real mentions on Reddit, Indie Hackers, Quora, niche forums
Reviews, case studies, and “we actually used this” content
AI trusts context, not slogans.
How to think about metrics now
Are you showing up in AI answers?
How often are you mentioned or cited?
Is the sentiment positive?
Are any signups coming from AI tools?
Page rank matters less. Visibility matters more.
TL;DR
Don’t just optimize for Google.
Optimize for where AI reads the internet.
Indie makers who adapt early get free distribution.
Everyone else gets summarized away.
This reframes the game really well. The shift from "ranking for keywords" to "being cited in answers" is fundamental.
One pattern I've noticed: AI models seem to heavily weight specificity and context over generic marketing language. A page that says "best project management tool" gets ignored, but one that explains "how we reduced sprint planning time from 3 hours to 45 minutes using X workflow" gets picked up because it reads as genuine experience rather than positioning.
The Reddit/IH/forum mention point is underrated. These models are trained on conversational data where real people discuss real problems. If your product gets organically mentioned in those discussions, you're essentially getting peer-reviewed by the training data.
Curious about measurement - have you found reliable ways to track AI-driven discovery? The attribution is murky since users don't always know (or say) they found you via ChatGPT or Perplexity.
Exactly. AI doesn’t reward claims, it rewards procedural specificity. “Best tool” is meaningless to a model; “here’s what broke, what we changed, and the result” is reusable reasoning.
Forum mentions work the same way. Repeated, problem-level mentions across Reddit/IH act like consensus signals, not promotion.
On measurement: we’re pre-analytics. I rely on prompt testing across tools, referrer hints where available, and self-reported attribution as directional signals. When organic forum mentions rise, AI visibility usually follows a few weeks later.
Feels a lot like early SEO before the tooling existed.
"Procedural specificity" - that's the exact framing I was reaching for. AI models can't do anything with "best" but they can pattern-match against "here's what we tried, what broke, and why this worked."
The early SEO parallel is apt. We're at the "figure out what works by experimenting" phase before proper tooling emerges. The prompt-testing approach makes sense - systematically checking if your product shows up in AI responses to relevant queries is basically manual SERP tracking for a new interface.
The forum-mentions-leading-AI-visibility lag you mentioned is interesting. It suggests the relationship is causal rather than just correlated - forums are literally training signal, not just proxy metrics. That changes how you think about content strategy: you're not just building brand awareness, you're contributing to the dataset that shapes future AI responses.
One experiment I've been curious about: intentionally using distinct, memorable phrasing when describing your product in forums, then tracking whether those phrases eventually appear in AI responses. Like tagging your own data to see if it gets picked up.
Yes....procedural specificity is the currency AI actually “spends.” Generic claims like “best tool” have zero utility because models can’t verify them or turn them into reasoning. What matters is concrete, reproducible steps: what failed, what was changed, and what the outcome was. That’s the pattern AI can cite and generalize.
Forum mentions function the same way. When multiple users discuss real problems and solutions across Reddit, Indie Hackers, or niche communities, it creates a consensus signal that models learn from. It’s not promotion it’s data that informs the AI’s reasoning.
On measurement, we’re still in the pre-analytics era. I use prompt testing across tools, track referrer hints where possible, and collect anecdotal attribution. Typically, there’s a lag: increases in organic forum mentions show up in AI visibility a few weeks later. It’s very reminiscent of early SEO before proper tracking tools existed.
One interesting experiment I’ve been running: deliberately using unique, consistent phrasing in forum posts almost like tagging your content and then checking if those phrases surface in AI responses later. It’s a kind of manual dataset engineering, but early results suggest it works.
The unique phrasing experiment is clever - essentially creating linguistic fingerprints to track your content's propagation through training data. It's like SEO keyword tracking but for a system where you can't inspect the index directly.
"Manual dataset engineering" is an honest framing. We're all doing this whether we admit it or not - every forum post is a potential training sample. The difference is doing it intentionally vs accidentally.
Curious what patterns you're seeing in the early results. Are certain types of phrasing more "sticky" than others? I'd guess concrete, jargon-adjacent terms (specific enough to be unique, generic enough to be reusable) would propagate better than purely invented vocabulary.