You launched something new: how long until AI notices?

No one can give you a reliable number of days. An assistant that searches the web can find a new page soon after it is public and crawlable, but being found is not the same as being recommended, and anything an assistant remembers from training updates far more slowly. Plan in weeks to months, measure repeatedly, and focus on removing obstacles.

The useful question is not "when" but "what is in the way".

Why is there no fixed delay?

Several separate things have to happen, and each has its own timing.

1. The page must be public and reachable. A crawler cannot read what it is blocked from or cannot find.
2. It must be discovered and fetched. Search indexes and assistants' retrieval work differ, and none publishes a schedule you can count on.
3. It must be relevant to a question someone asks. A page nobody asks about is never retrieved.
4. It must be judged worth using. Even a found page competes with others already trusted.
5. The answer must vary your way. Recommendations differ between runs, so you may appear in some answers before others.

That is why a launch can be visible in one assistant and absent in another for a long time. For the mechanics of where answers come from, see training data or live search.

What is the difference between being found and being recommended?

A new product can be found, named neutrally in a few answers, and cited from its own site long before it is recommended. Recommendation tends to need more: independent descriptions of you on other pages, a clear fit with the question and enough evidence for the assistant to present you as a choice.

This is why it helps to track several things separately: whether you are named at all, whether your site is cited, and whether you are recommended. Mentioned, recommended or cited explains the differences. A young brand often moves on the first two before the third.

What does memory have to do with it?

Assistants carry knowledge from training, and that updates in large steps when new versions are released, not day by day. A product launched recently may not be in that built-in knowledge at all. When an assistant does not search the web for a question, it may know nothing about you, or only the older picture.

Assistants that search the web can fill that gap, which is why the live path matters for new things. You cannot control whether a particular answer uses search. You can make sure that when it does, your information is there to find.

How do you make it easy to be found?

Remove obstacles, in this order.

• Make the page public. No login wall, no blanket noindex, no staging URL left in place.
• Allow the crawlers. Check robots.txt and any bot protection. See AI crawlers and robots.txt.
• Link to it from your own site, from the homepage or a relevant hub, not only from a deep page.
• Submit a sitemap so search engines learn the URL.
• Write the page clearly. What it is, who it is for, the facts. See why your homepage confuses AI assistants.
• Update existing pages that describe your old offering, so the new and the old do not contradict each other.

How do you give assistants more to work with?

A launch announced only on your own site is one voice. Independent mentions add weight and also give assistants other pages to cite.

• Announce it where your buyers read: relevant communities, newsletters, trade press, partner pages.
• Brief reviewers and roundup editors honestly. How to get onto the roundup and review pages AI relies on covers the approach.
• Make it easy for early customers to say what they think, openly, as in reviews, forums and community answers.
• Keep the story consistent. The same name, description and facts on every page reduce confusion.

How do you tell whether AI has noticed?

Measure, do not guess.

1. Choose the questions a buyer would ask where your launch should be relevant. No brand names in them.
2. Add a few branded questions such as "what is [product]?" to see whether it is known at all. Keep these separate, because they answer a different thing.
3. Ask all four assistants, signed out, several times, and save the answers.
4. Record whether you are named, in what role, and which of your pages are cited.
5. Repeat on a regular schedule using the same questions.

Answers vary, so wait for a pattern before reading anything into it. How many times to ask and how to track AI visibility over time give the routine.

Does a bigger launch get noticed faster?

Not reliably. A launch with wide coverage gives assistants more pages to find and cite, so it can help. But coverage that is thin, repetitive or promotional adds little. A smaller launch with a clear page, an accurate description and a few credible independent mentions can do as well. The quality and consistency of what is published matters more than the volume.

What should you do while you wait?

Keep working on the things you control: the clarity of your pages, the accuracy of what others say about you, and the evidence that real customers use and value the product. Those matter whatever any assistant does. If a launch seems to have been ignored for a long time, go back through the checklist above before assuming something deeper is wrong.

To see how a site appears to assistants at a given point, try the Citewake checker. For how Citewake treats a brand with no trace in its benchmark, including the difference between "no evidence" and "not indexed", see the methodology.