How long does AI SEO take to show results?

What do vendors say the timeline looks like?

AirOps, a vendor, gives the clearest published timeline. The first month is setup, and it says you should not expect visibility changes yet. Well-structured pages on authoritative domains start to appear in AI answers between day 30 and 60, with mentions as the first signal. Citations follow between day 60 and 90 as engines crawl and verify the content again. It says the timeline depends on existing domain authority, content structure and measurement discipline¹.

An Indian agency's benchmark describes the same shape with different dates. AI visibility lags 30 to 60 days after content publishes, first citation wins arrive between day 45 and 90, and a 3x to 5x rise in citation share on primary prompts takes 6 to 9 months. The same page says retainers usually carry a six-month minimum².

Both firms sell the service, and neither publishes a sample size or a method for its dates. AirOps says only that its teams have tracked thousands of pages¹.

Is there independent evidence on timing?

We did not find an independent study of how long AI answers take to change. That is a gap in the evidence. It is a reason to measure your own site from a baseline rather than a reason to wait.

Classic search data helps set expectations. Ahrefs found that 1.74 percent of one million new pages reached Google's top 10 within a year, down from 5.7 percent in 2017. It also reports 6.11 percent in a second sample of two million English-language pages created in October 2023, which it describes as a fairer figure³.

AI answers add another source of noise because they change between runs. Commenters in one Hacker News thread noted that the same prompt can return different recommendations from day to day⁵. AirOps reports that only 30 percent of brands stay visible from one answer to the next, and that about 57 percent of those that disappeared resurfaced within two runs. Both figures are the vendor's own⁴. A flat week is not proof of failure, and a good day is not proof of success.

What should I do in week one?

Build a baseline before any work starts, so that day 60 has something to be compared against.

• Write the question set: 20 or so questions that real buyers ask, drawn from sales calls, support tickets and Search Console queries. AirOps names those same sources⁴.
• Run them logged out in the assistants your buyers use. Record whether you were named, cited or missed, and how you were described.
• Fix the cheap blockers first: crawler access, wrong or missing facts, and pages that do not answer the question.
• Repeat on a schedule and read windows, not single runs. AirOps suggests windows of 7 days or more, weekly checks as the minimum for decisions, and a quarterly reset of the question set⁴.

What if the numbers are flat after 60 days?

AirOps says that if metrics stay flat past day 60, teams should revisit targeting, structure, internal linking or schema. It also lists JavaScript-heavy pages and content older than six months among the factors that slow results¹. Put reviews at day 60 and day 90 into the contract. One contract guide recommends hard KPIs with formal reviews at those two points⁶.

Measure the click side too, but expect gaps. Google says AI Overviews and AI Mode traffic is reported in the Search Console Performance report under the Web search type⁷. A newer Generative AI performance report shows impressions only, with AI Overviews and AI Mode combined, and was reported to be live worldwide as of 31 August 2026⁸. Some AI-influenced visits arrive as direct or organic traffic, according to commenters in the Hacker News thread⁵.