Get started

AEO checklist

Use this checklist when you start AEO work, update content, or report results to a client. Keep your questions and observation conditions consistent, then use the evidence in AI answers to plan improvements and the next review.

Click an item to mark it as reviewed. Checks reset when you reload the page. Record evidence, owners, and deadlines in your team's work management tool.

Agent prompt

Use https://orchestor.io/docs/en/aeo-checklist to review our AEO work.
Confirm our brand, domain, audience, region, language, AI services,
question list, observation period, and goals.
Mark each item as verified, needs work, unverified, or not applicable.
Include the supporting answer, URL, and review date.
If you cannot access the data, mark it unverified and ask for what you need.
For each improvement, specify the URL, owner, completion criteria,
and next review date.
For agency work, confirm responsibilities, publication approval,
and report recipients with the client.

Goals and scope
Baseline observations
Search and crawling
Information accuracy
Making improvements
Re-observation and reporting
Agency operations

Goals and scope

Define your audience, products or services, target region, and language. Record the business goal you want AEO to support.
Check your brand and competitors, including names, aliases, and domains, so comparisons use the intended entities.
Select prompts your customers actually ask. Include branded and unbranded questions across research, comparison, and selection. Organize the sample with topics and tags and remove duplicates.
Choose the AI services and number of questions to observe. Check your plan and usage allowances.

Baseline observations

Record the baseline question list, AI services, region, language, and period. Keep later changes to these conditions distinguishable.
Confirm that AI answers are being collected. Record the answer count and any failures or missing observations.
Review the metric definitions. Record brand mentions separately from website citations, and compare competitors under the same conditions.
Open representative answers and their cited URLs. Check for mistaken brand identities, inaccurate descriptions, and differences from competitors.

Search and crawling

If Google is in scope, use Search Console or an equivalent check to confirm that key pages are indexed and eligible to appear with a search snippet.
Check whether robots.txt, your CDN, or authentication unintentionally blocks the selected service's search crawler. Decide on search access separately from model-training access.
Provide important explanations as readable text, not only in images. Confirm that internal links from relevant pages lead to this content.
If you use structured data, check that product names, prices, availability, and organization details match what the page displays.

Google does not require special AI files or dedicated structured data for AI Overviews or AI Mode. Start with its standard search requirements. ChatGPT's search crawler, OAI-SearchBot, and training crawler, GPTBot, can be configured independently. Allowing access does not guarantee inclusion or citation.

Information accuracy

Identify an existing page that answers each priority question. Where no answer exists, specify the information to add and its target page.
Verify brand names, product specifications, prices, and terms against the facts. Correct contradictions and outdated statements across your website.
Support numbers, comparisons, and specialist claims with evidence that readers can verify. Include sources, methods, authors or reviewers, and update dates where relevant.
Read cited pages to identify explanations missing from your own pages. If a third-party description is wrong, record where to request a correction and the supporting evidence.

Making improvements

Use answers and citations to select the first issues to address. Record their effect on customer decisions and the work required to fix them.
Give each improvement a target URL, proposed change, owner, deadline, and a way to verify completion.
Verify the facts and obtain the required approval before publishing. Keep a record of the previous version and what changed.
Open the published page and check the updated content, links, viewing access, and crawl access. Record its public URL and publication date.

Re-observation and reporting

Set the next review date and compare performance using the same questions, AI services, regions, and other conditions as the baseline. Disclose changes to those conditions in the report.
Check metric changes against answer text and cited sources. Do not attribute improvement to your changes based only on a single answer or a few observations.
Report AI mentions and citations separately from visits measured in analytics and pipeline or revenue measured in a CRM. Mark outcomes you have not measured as unverified.
Record findings, remaining issues, next actions, owners, and the next review date. Preserve comparison context when you revise the questions or competitors being tracked.

Agency operations

When supporting a client, agree on the following operating conditions in addition to the shared checklist.

Agree on the brands, scope, deliverables, and exclusions. Keep each client's observations and work records separate.
Assign responsibility for account access, fact checking, and publishing. Record the client and agency contacts.
Share the baseline conditions, results, and metrics used for evaluation. Agree on report recipients, reporting frequency, and approval before publication.
Communicate delays and scope changes. Include supporting answers, URLs, and both parties' next actions in reports. Agree on access removal and handover when the engagement ends.

​AEO checklist

​Goals and scope

​Baseline observations

​Search and crawling

​Information accuracy

​Making improvements

​Re-observation and reporting

​Agency operations