Discoverability
I launched my SaaS and nobody is finding it. What now?
A quiet launch is the normal start for a young product. What our AI search data shows, the three checks to run in order, and the one move to make after them.
A mention and a citation tell you different things.
Which product fits my team?
A recommendation names a product and may point to a source page ↗.
- Mention
- Who was named?
- Citation
- What page was linked?
In this article
If you launched a few weeks ago and the sign-ups have not come, you are where almost every young product starts. Search engines and AI assistants have little to say about a product nobody has written about yet. The useful move is to find which step is failing, fix that one, and leave the rest alone for now.
A quiet launch is the normal starting point
In our cohort study, young Y Combinator companies were named in 6.4% of usable AI search answers, against 36.7% for YC's graduated companies1. 48 of 118 young companies were never named in the answers about what they sell. In a separate scan, ten companies from recent Y Combinator batches had a median mention rate of 0.8%, against 83% for ten household names2.
These are companies Y Combinator chose to fund. Being absent from the answers was still the default for them, so a quiet start says little about your product on its own.
What the engines told a founder asking this exact question
We track this question every week. On 7 October 2026 none of the four engines named a tool in its answer, and they did not agree on where to start3.
- ChatGPT, through the OpenAI API with web search, started with indexing: check that search engines have your pages, then submit a sitemap.
- Gemini started with the niche: name the pain you remove and the first 100 people who feel it.
- Google AI Overviews started with people: stop waiting for organic traffic and talk to 20 potential users.
- Perplexity started with measurement: track visitors, sign-ups and first use before changing anything.
Each answer is right about one failure. A founder whose pages are not indexed gets nothing from 20 conversations about positioning, and a founder whose page confuses every visitor gets nothing from a sitemap. Checking in order tells you which of the four is talking about your product.
Check these in order
1. Check that search engines can reach the page
Open Search Console and inspect the address of the page you most want people to find. It tells you whether Google has indexed the page and, if not, why4. If the answer is no, fix that before writing anything new, because nothing else on this list works on a page that is not in the index.
2. Check whether anyone sees it
In the Search results report, look at impressions and the queries behind them5. Few impressions on relevant queries means people are not being shown the page. Impressions on unrelated queries means the page is not saying clearly who it is for.
Then ask ChatGPT, Gemini, Google AI Overviews and Perplexity the questions your customers ask, and write down who is named and which pages are cited. Google says there are no additional requirements to appear in AI Overviews or AI Mode, and no special markup to add6. Save the date beside each answer. A missing mention on one day is a starting reading to compare against later.
3. Check whether the people who see it understand it
If people do arrive, watch one person who fits your audience try the page without help. Can they say what the product does, and reach a first useful result? A page that attracts visitors and then confuses them needs different work from a page nobody sees.
4. Pick one move and finish it
Choose the move that fixes the step that failed: an answer in a community thread where your product genuinely helps, with your connection disclosed; a page that answers a question your customers ask; a complete profile on a directory the engines already cite. Finish it, note the date, and take the same reading again a few weeks later.
Where first users actually look, and how discovery compounds →How to get your startup cited by AI assistants →What to leave until later
- Paid ads, until the visitors you already get turn into sign-ups.
- Submitting to dozens of directories at once. A few complete listings on sites the engines already cite are a better start than a long list.
- Another feature. Until you know which step is failing, a new feature arrives with the same problem as the old ones.
Where AfterLaunch fits
AfterLaunch runs these checks continuously. It reads your site, asks the leading AI search engines the questions your customers ask, watches search, communities and directories, and ranks the next move with its evidence and a working draft in your voice. You or your AI finish it, and nothing ships without your approval. When we asked this article's question on 7 October 2026, no engine named AfterLaunch either3.
- AfterLaunch, One answer in sixteen: how often AI search names a young company, 25 August 2026.
- AfterLaunch, We scanned ten companies nobody has heard of, 22 August 2026.
- AfterLaunch weekly scan of the question "I launched my SaaS and nobody is finding it, what now" on ChatGPT (OpenAI API with web search), Gemini, Google AI Overviews and Perplexity, read on 7 October 2026. Our own record, not published separately.
- Google, URL Inspection tool, Search Console Help. Checked 8 October 2026.
- Google, Performance report (Search results), Search Console Help. Checked 8 October 2026.
- Google Search Central, AI features and your website. Checked 8 October 2026.
The free scan reads your site, then asks the engines about you. No card.
Scan my siteRead next
What sources do AI answer engines cite?
Where AI answer engines find their sources: observed citation patterns, measured separately by engine, sample and counting method.
Read →JournalHow Reddit citations differed across AI search engines
Reddit citation rates varied sharply across our matched question set. The study also shows why the OpenAI API and ChatGPT Search need separate labels.
Read →