AEO Consulting: What Should Arrive in the First 90 Days?
Instead of chasing vendor recommendation lists, judge AEO consulting by the deliverables you actually hold after 90 days.
Search for “AEO consulting recommendations” and you mostly get lists of vendor names. But six months later, satisfaction usually depends less on which firm you picked and more on what deliverables you actually hold after 90 days. Writing down the outputs you expect — before you read any list — makes comparison far easier.
What should you define before shortlisting vendors?
Define your question set and your baseline first, not your vendor. AEO is about whether your brand gets cited in AI answers, and that outcome shifts entirely depending on which questions are asked. Writing 30–50 questions up front turns vague proposal language into a verifiable promise.
- Question set: the phrasing a real customer would type into ChatGPT or Gemini, such as “best orthodontic clinic in Gangnam.” Split them into branded and unbranded queries.
- Baseline capture: run those questions yourself before signing and save the raw answers. Without this, you lose any basis for judging improvement later.
- Ownership: who keeps the structured data, content, code, and accounts when the engagement ends. This belongs in the contract as a written clause.
In what order should deliverables arrive?
A healthy engagement stacks up as diagnosis → build → measurement. If month one consists only of publishing content with no diagnostic document, you may be paying for volume without evidence.
- Weeks 1–2 · Diagnostic report: how crawlers and AI systems currently see your site — JSON-LD, robots, sitemap, rendering method — plus measured results for your question set.
- Weeks 3–6 · Build: structured data, answer-shaped content (question headings with a direct answer paragraph), AI-facing files such as llms.txt, and internal link cleanup.
- Weeks 7–12 · Measure and iterate: re-run the same question set, compare before and after, analyze why some pages get cited and others don't, and produce the next quarter's backlog.
Common mistakes in this window
The most common mistake is publishing repeatedly without measuring. Next is declaring success or failure from a single day of AI answers, which ignores how variable they are. Third is bolting AEO onto a site whose basic search foundations are broken. If the terminology is still fuzzy, align definitions first with our complete SEO, AEO and GEO guide.
How can you inspect the deliverables yourself?
You don't need to be a specialist to review this work. Many items can be checked with a browser and public tools alone. Use the list below at every monthly report.
- View page source and confirm the body text actually exists in the HTML — content painted only by JavaScript may not be read.
- Check that JSON-LD parses without errors and matches what the page actually says.
- Re-ask the same question in a fresh session and on a different model to see whether the result tends to reproduce.
- Confirm the report includes raw answer screenshots with dates. Score-only reports are hard to verify.
- Confirm the work lives in your repository and your accounts.
How does SAENRU approach this?
To be transparent about our own service: SAENRU runs integrated SEO, AEO and GEO consulting, covering Google and Naver search alongside citation in ChatGPT, Gemini and Claude answers.
- Measured diagnostic report: KRW 190,000 over 5 business days — a way to see your current state in numbers and raw answers first.
- Build package from KRW 3,600,000 and monthly operations from KRW 1,470,000 per month. The diagnostic fee is fully credited if you proceed to a build contract. All prices exclude VAT.
The SAENRU site itself is an AEO reference build: raw HTML with embedded JSON-LD and llms.txt, so you're welcome to open the source before you commit. AEO results are not settled in a few weeks, so starting small with a diagnosis and deciding on expansion after reviewing 90 days of deliverables is usually the sounder path.