seoaraci.com
SEO tooling product we built; the measurement mindset behind our audits started here.
seoaraci.com
Read the case study →AI Search · Audit and technical layer
AI Search Optimization
AI search optimization is the technical and diagnostic side of being found by ChatGPT, Gemini, Perplexity, Copilot and Google AI Mode: can the crawlers reach your pages, is the important content in the HTML, do the structured data and entity signals say the same thing everywhere, and how often do answer engines actually mention you today. POI369 starts with an AI visibility audit, then fixes the technical layer and sets up monthly measurement.
Last reviewed September 14, 2026 · Reviewed by Ahmet Bulut
Audit first, then fix, then measure
Most companies have no idea whether AI assistants mention them, and if they do, from which source. Guessing is expensive: teams rewrite pages that were never the problem while a robots rule silently blocks the one crawler that mattered. So the service starts with facts.
The audit answers four questions. Access: which search and AI crawlers can fetch the site, and what do they get back. Content: which pages carry the brand’s answers in server-rendered HTML, and which hide them behind scripts, tabs or images. Signals: do the Organization, Person, Service and Article data, the canonical URLs, hreflang and sitemap describe one consistent entity. Visibility: for a fixed set of prompts, how often is the brand mentioned or cited, where, and next to whom.
The output is a written report with a prioritised fix list. Some fixes take an hour. Some are a content program, which is where our GEO service continues.
What the audit covers
Crawler access
robots.txt rules for Googlebot, Bingbot, GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, Google-Extended and others. Content-Signal declarations. Firewall and bot-management rules that block by accident.
Rendered vs. delivered HTML
What a crawler receives without executing JavaScript, compared with what a browser shows. Missing headings, intros, service descriptions and Q&A get flagged.
Structured data and entity graph
Validity is the easy part. We check that the data is truthful, connected by @id, consistent with the visible page, and free of fabricated ratings or partner claims.
Canonical, hreflang and sitemap
Self-canonical language versions, alternates that point only to pages that exist, lastmod values that change when content changes.
Freshness signals
IndexNow for Bing and partners, sitemap hygiene, visible review dates. Bing’s AI Performance report is set up so AI-driven impressions become visible.
Baseline visibility benchmark
Thirty to fifty prompts, brand and category, run across ChatGPT, Gemini, Perplexity and Google AI Mode. Mention rate, citation rate, position, competitors, source URLs.
Typical findings
Patterns we see repeatedly. None of them are exotic; all of them cost visibility.
AI crawlers blocked by a default rule
A bot-management preset or an old robots.txt denies GPTBot or PerplexityBot. The brand cannot be cited from a page that was never read.
Answers hidden in JavaScript
The service description loads from an API after render. The crawler sees a shell. Moving that content into server-rendered HTML fixes it.
Three spellings of the company
Brand name, legal name and directory listings disagree on casing, suffix or address. Engines hedge on ambiguous entities.
Schema that says nothing
Organization data with a name and a logo and no description, founder, services or sameAs. Valid, and nearly useless.
No evidence pages
Claims of expertise with no case studies, no author, no dated content. Engines prefer sources that show their work.
How the engagement runs
Week 1: access and content audit
Crawl the site as the engines do. Compare delivered and rendered HTML. Review robots, sitemap, canonical, hreflang, structured data.
Week 1 to 2: visibility baseline
Run the prompt set across the four engines, record everything in a shared sheet. This becomes the reference for every later month.
Week 2 to 3: fixes
Technical fixes shipped by us or handed to your developers with exact instructions. Entity data corrected. IndexNow and Bing AI Performance configured.
Monthly: measurement
The benchmark runs again. You get the numbers, the deltas and a short note on what to do next. Continue into a GEO program if the content side needs work.
Related work
karasuemlak.net
Real estate listings platform where local search visibility depends on crawlable structured pages at scale.
karasuemlak.net
Read the case study →poitim.com
Five-language SaaS product: an example of the canonical, hreflang and entity discipline we audit for.
poitim.com
Read the case study →Questions we get asked
Is this a one-off or a subscription?
The audit and fixes are a fixed-scope project. Monthly measurement is optional and priced separately. Many clients continue into the GEO program once the technical layer is clean.
We already rank well on Google. Do we need this?
Ranking is a good sign, since answer engines draw heavily on search results. But they also weigh entity clarity and whether they can lift a clean passage from your page. The audit tells you in two weeks whether anything is in the way.
Can our own developers apply the fixes?
Yes. The report lists each fix with the exact file, rule or field to change. We can also ship them ourselves on Next.js, WordPress and most common stacks.
Which tools do you use to measure?
Google Search Console, GA4, Bing Webmaster Tools with the AI Performance report, IndexNow, and a manual monthly benchmark across ChatGPT, Gemini, Perplexity and Google AI Mode. We do not scrape the engines in ways that break their terms.
Related expertise and reading
Related expertise
Related insights
How AI Search Discovers, Understands and Cites a BrandBetween a company’s website and a sentence in a ChatGPT answer there are four stages: crawling, retrieval, entity resolution and citation selection. Each one can silently drop a brand. This is a plain-language map of the pipeline and of what a brand controls at each stage.How We Structure Next.js Websites for Search and AI CrawlersThe concrete decisions behind poi369.com and the products we build: what goes in server-rendered HTML, how a page serves Markdown to agents on request, why sitemap dates are derived from content signatures, how six languages stay self-canonical, and how structured data forms one graph. Everything here is running in production.How to Improve Brand Visibility in ChatGPT and AI SearchA working checklist, in the order we apply it: measure first, open the doors, make the entity unambiguous, write passages an engine can lift, publish evidence, get corroborated, measure again. No tricks, no guarantees, and every step is something you can verify yourself.
Related work
seoaraci.com
Tool and automation platform for SEO and content teams.
WebAIPlatform
karasuemlak.net
Real estate listing platform for the Karasu region with an AI-powered content engine.
WebAIPlatform
poitim.com
Global B2B team management platform, built end to end from brand identity to the desktop app, in five languages.
SaaSPlatformAIBrandWebMulti-lang
Get the audit
Send your domain and your main markets. We reply with the audit scope, timeline and a fixed price.