There is no reliable, universal number of weeks or months at which AEO "starts producing results". When you see measurable change depends on the business's starting point, whether its content is accessible and discoverable by the relevant systems, the quality and consistency of its public information, and how each platform selects sources for a specific question. At Praion, we use 30, 90 and 180 days as checkpoints — not as promised result dates.
First define what "results" means
"Results" can mean different things.
You may want to see whether:
- important pages are accessible and discoverable by the relevant systems,
- the business's public information is more complete and consistent,
- the business or one of its pages starts being mentioned in relevant answers,
- the same mentions recur across successive measurements,
- the right public sources are being used,
- errors or gaps in the information being surfaced are reduced.
These things do not necessarily happen at the same time.
The fact that a technical change has been implemented and can be discovered does not mean the page will be indexed by a search engine, appear in a result or be used as a source in an AI answer.
That is why a blanket rule such as "AEO takes three months" is not precise enough.
What do you check in the first 30 days?
The first checkpoint is mainly about the foundation and establishing a baseline.
You examine:
- whether important pages are technically accessible and do not block crawlers that need access through robots.txt, a CDN or other infrastructure,
- whether, where required, search engines can discover and index them,
- whether core content is available as text,
- whether information about the business, its services, products or location is accurate,
- whether the same core details are consistent between the website and other public sources,
- whether structured data — the page's technical markup — describes real, visible content,
- what the baseline looks like across the questions and systems you want to track.
That last point is critical. If you have not recorded the baseline, it will be much harder to know later what actually changed.
The technical requirements are not identical across platforms. In Google, for example, a page must be indexed and eligible to appear in Search to be eligible as a supporting link in AI Overviews or AI Mode. In ChatGPT Search, access for OAI-SearchBot affects whether content can be discovered and surfaced.
Google does not require a separate "AI schema" for AI Overviews or AI Mode. Structured data remains useful when it accurately describes visible content, but no schema can guarantee appearance, mention or recommendation in an AI answer.
What makes sense to look at around 90 days?
The 90-day mark is a comparison point, not a general date when "results arrive".
You repeat the measurements over a fixed set of questions a real customer could plausibly ask, and record:
- whether the business or its content is mentioned,
- for which questions,
- by which system,
- which source is used when a citation is provided,
- whether mentions recur or appear only occasionally,
- whether inaccurate or outdated information is being surfaced.
For example, if you are tracking an accounting firm, a realistic question could be:
"I have a sole proprietorship and want to switch accountants. Which accounting firms in my area publicly state that they serve small businesses, and what information do they provide before I get in touch?"
For a question like this, an AI system that uses public sources may, depending on how it works, draw on information from the firm's website, its Google Business Profile, public professional profiles or other content published by the business itself.
It may surface or summarise facts that are publicly documented, such as location, contact details or services. If an answer infers that the firm serves a specific type of business, has availability or is the "best choice" without public evidence, that inference should not be treated as reliable.
A single mention is not enough to establish a trend.
The comparison is more useful when it is repeated with the same questions under conditions that are as similar as reasonably possible. This helps distinguish an occasional result from a pattern that recurs.
What do you evaluate after 6 months?
At six months, we do not assume that a "stable presence" has been created.
But if you have measured consistently from the start, you now have more data with which to compare the current state against the baseline.
You examine:
- whether more questions produce recurring mentions,
- whether the sources being used are the right ones,
- whether significant gaps or incorrect information remain,
- which content needs updating,
- whether the website and other public sources are consistent with one another,
- which parts of the work are worth continuing or changing.
The goal is not to reach an arbitrary number of appearances.
There is no universal rule saying that "after six months a business should appear for X questions".
What can delay or limit progress?
Before concluding that AEO "isn't working", check whether there is a specific technical or informational problem.
For example:
- important pages are blocked from crawlers that need access or, where required, cannot be indexed by a search engine,
- public information is incomplete or contradictory,
- core content exists only in images or in formats that are harder for the relevant systems to process,
- important pages are not linked to one another correctly,
- structured data does not match the visible content,
- outdated information remains published,
- the content does not answer the audience's real questions.
Even when all of these are in order, there is no guarantee that a search engine or AI system will use a specific page or business as a source.
If you are not sure what AEO actually includes, start with what AEO is.
So, how much time should you allow?
For planning and measurement, we suggest three practical checkpoints:
- 30 days: technical foundation, public information and baseline measurement,
- 90 days: first organised comparison of repeated measurements,
- 180 days: evaluation of trends, gaps and next priorities.
These checkpoints are Praion's tracking framework. They are not an external benchmark and do not predict when an AI system will display, mention or use a business as a source.
Their value is that they create a consistent way to compare progress instead of relying on a single good or bad answer.
If you want us to look at what can be measured in your own case, see our approach or let's talk.
AEO work can help make a business's public information more complete, accurate and consistent.
It does not guarantee appearance, mention, citation, recommendation, a specific placement or a specific business outcome in ChatGPT, Gemini, Perplexity, Google AI Overviews or any other system.