How a one-person app gets found by AI assistants
In October a Thuuz user wrote to me. Thuuz was the app that told you which games were worth watching, and it has been gone since early 2021. He'd asked three AI assistants for a replacement. Only ChatGPT came back with The Lineup Is. The other two had never heard of it.
That message changed how I think about being found. People don't only search anymore. They ask an assistant for "an app that tells me which games are worth watching" and take whatever it names. If you're one person with one app, you need to know where those answers come from. Here's what I've learned so far. We're early, and most of this is still in progress.
Each assistant reads a different index
The first useful thing I learned was that "AI search" isn't one thing. When an assistant looks something up, it goes to a search index, and they don't share one.
- ChatGPT leans on Bing, plus its own crawler. When it found us, it cited our homepage.
- Gemini leans on Google.
- Claude's web search uses Brave.
That explained the Thuuz user's result right away. We'd done the Bing setup, so ChatGPT could see us. Google had crawled the site but indexed very little of it. And when I checked Brave, it hadn't picked up our site at all. Claude couldn't recommend a page its search engine didn't have.
So the job isn't "do SEO." It's three or four separate jobs, one per index, and each one has its own way in.
Bing: verify once, then push
Bing was the easiest. Bing Webmaster Tools can import your site from Google Search Console. That import is a one-time snapshot, so nothing needs re-adding later. After that, IndexNow is Bing's push channel: you tell it when a URL changes and it comes and looks.
Our site has pages that change every day, so the server pings IndexNow with the day's static pages automatically. One gap I found: new blog posts aren't in that daily batch, so I ping those by hand.
Brave: it doesn't take IndexNow, but it has a form
This one surprised me. Brave isn't an IndexNow participant, and it has used only its own index since April 2023. Everything I'd pushed to Bing never reached it.
Brave documents two ways in: its own crawler, which follows links from pages it has already crawled, and the opt-in Web Discovery Project. It also has a submit form at search.brave.com/submit-url, which is reported to work for new URLs. Getting linked from pages Brave already knows matters too. A site that auto-lists App Store apps already linked to us, which helps.
A month later Brave has started picking up our pages. When I search Brave for the category, things like "Thuuz alternative app" or "app that tells you which games are worth watching", I get Wikipedia, tech press and other apps' App Store pages. Not us yet. Searching our own name surfaces a betting app with a similar name first. That's the honest state of it.
Google: authority, not discovery
Google found our pages fine. Most of them sat in "Discovered, not indexed" or "Crawled, currently not indexed." A full re-crawl of the site showed the technical side was clean. That status, as far as I can tell, is about authority. Google knows the pages exist and doesn't think they're worth keeping yet.
I also want to save other founders some time on llms.txt. We have one, a plain summary of what the app is and what every page covers. But Google has said it doesn't read llms.txt. I keep ours up to date because it's cheap, and I don't invest more than that.
What I did do is make sure our robots.txt welcomes the AI crawlers by name, OpenAI's, Anthropic's and Perplexity's among them, alongside Bing's and Google's. Blocking them by accident is an easy way to stay invisible.
What helped
Assistants summarize. They pull from listicles, Reddit, YouTube, directories and pages that answer the question directly. So most of my work has been building pages an assistant can lift a clean answer from.
Pages that answer the exact question in plain sentences
If someone asks for an app that alerts them when a game gets close, the best page to have is one that says, in the first sentence, that this is an app that alerts you when a game gets close. Our close-game alerts page does that, then explains what counts as close in each sport. No clever headline, no brand voice in the way.
A page per question
Early on I had one homepage trying to answer everything. Now each question people actually ask gets its own page: watching sports without spoilers, scores on the lock screen, close-game alerts. Every finished game also gets a page that answers "was it worth watching?" in one spoiler-free line.
FAQ markup
The question-shaped pages carry FAQ structured data, so the question and its answer are marked up as a pair. It costs almost nothing to add, and it forces you to write an answer that stands on its own.
Comparison pages
People ask "is X better than Y?" all the time. We have comparison pages against the apps people already use for scores. Two rules I hold to: every fact about another product is checked against that product's own listing, and we never quote another company's prices, because those change and a stale price is a false claim about someone else.
A replacement page for a dead product
This is the one I'd tell every indie founder to look for. Thuuz shut down years ago and people still search for it. Our Thuuz replacement page says plainly what happened to Thuuz, when, and what does the same job now. Those searchers are the warmest audience we have, and assistants are where they ask. If a product in your category died, write the page that answers "what do I use now?"
Day pages
"What games are worth watching tonight" is a question that changes daily, so the answer has to as well. /tonight ranks the current slate and updates live. There's a page for every sport and league, a page for each upcoming day, an archive for every past day, and /this-weekend for people planning ahead. I put that one at its own URL instead of under /tonight, because the app already claims /tonight links and would have opened the wrong screen.
Video
YouTube is one of the most cited domains across AI engines, so I've leaned into it: short, captioned videos made from the app's real screens, also embedded on the site with video structured data. One snag worth knowing: YouTube locks uploads from unaudited API projects to Private, and locked videos can't be made public. I route uploads through a service with an already-verified YouTube connection while the audit is pending.
What didn't help, or hurt
- Waiting. After the first big site overhaul in July, search still hadn't noticed us seven weeks later. Nothing indexes until you've verified the site and submitted a sitemap, and I hadn't done either.
- A broken share link. For a while every link the app shared pointed at a domain we didn't own and that didn't resolve. Every share was a dead end instead of a backlink. Check yours.
- Listing live pages in the sitemap. A page for a game in progress could change while Google was looking at it and drop out of the index. The sitemap now lists finished games only.
- Directory queues. Some directories are months behind on free submissions. A few offer paid priority review. Plan for the lag.
llms.txtas a strategy. As above, fine to have, not a lever.
Where we are
On-site, the gaps are mostly closed. The pages exist, each index has a way in, and day pages keep the site fresh. The real lever now is off-site: links from sites the indexes already trust, roundups, directories and videos. That part is slow, and I'm in the middle of it.
If you're a solo founder, here's the short version. Figure out which index each assistant reads. Get into each one through its own door. Then write the page that answers the question your users are typing, in the words they type it.
The Lineup Is is an iPhone app that ranks every game by whether it's worth watching.
Get The Lineup Is on iOS — free — or see tonight's best games ranked right now. More from the blog.