Technical field guide AI Search Readiness
What business owners should know
Your website may rank in Google and still be invisible to AI.
AI search visibility is not one trick, one plugin, or one score. Before ChatGPT, Perplexity, Gemini, Claude, Google AI, or Copilot can recommend a business, the site has to be crawlable, readable, extractable, trusted, and measured engine by engine.
Built from public crawler documentation, search fundamentals, and Alchemetics field testing.
The mental shift
SEO gets you found. AI search decides whether you are worth recommending.
Traditional SEO was built around ranking position and clicks. AI search adds another layer: citation and recommendation. A page can rank, get crawled, become a candidate source, feed part of an answer, and still never be named. That is why Alchemetics audits the whole path, not just the ranking.
Crawl access
Can the right bots reach the pages, or are robots rules, redirects, security tools, or bad status codes blocking them?
Readable HTML
Does the important content exist in the initial HTML response, or is it hidden behind JavaScript that some crawlers may not render?
Extractable answers
Does each page answer real buyer questions clearly enough that a passage can stand alone inside an AI answer?
Entity trust
Does the web clearly agree who the business is, what it does, where it operates, and why it should be trusted?
Citation proof
Are there credible third-party sources, reviews, directories, mentions, and proof pages that support the recommendation?
Engine measurement
Are ChatGPT, Gemini, Perplexity, Claude, Google AI, and Copilot tracked separately instead of blended into one vague score?
Step 1
Make sure AI crawlers can access the site.
The first failure is often silent. A robots.txt file may allow Google but block an AI retrieval bot. A firewall may challenge a crawler even when robots.txt says it is allowed. A redirect chain may work for a browser but fail for a faster crawler. The audit starts by checking access bot by bot.
Step 2
Make the page readable without relying on JavaScript.
Many modern sites look complete to humans but expose very little useful text in the raw HTML. That creates risk for AI retrieval systems, partner search indexes, and tools that do not fully render JavaScript. The safest approach is server-rendered or statically rendered content where the important words are present immediately.
The simple test
Fetch the page as raw HTML. Turn JavaScript off. Compare that text against the rendered browser page. If the headline, service copy, FAQs, proof, and internal links disappear, the site is not AI-search ready.
Use server-side rendering, static generation, clean HTML, crawlable links, schema, and content that exists before scripts run.
Step 3
Write pages so answer engines can extract the right passage.
AI systems do not always use a whole page. They retrieve sections, compare sources, and assemble answers. That means each important section should work by itself: direct answer first, clear entity names, specific proof, useful structure, and no vague filler.
Answer-first opening
Start the page with a direct 40 to 60 word answer to the main buyer question.
Question-shaped headings
Use headings that match how people ask AI tools for help, not only old keyword fragments.
Complete sections
Each section should make sense even if AI pulls it away from the rest of the page.
Proof density
Use dates, numbers, names, examples, reviews, case studies, and credible references.
Structured formats
Tables, lists, comparisons, FAQs, pros and cons, and best-for sections are easier to parse.
Freshness
Show current dates when pages are updated and refresh important answers regularly.
Step 4
Build trust beyond your own website.
Owned content matters, but AI engines often need corroboration. If competitors are supported by reviews, directories, Reddit threads, YouTube videos, trade publications, partner pages, and comparison lists, they may be easier to recommend. Alchemetics identifies those source gaps and turns them into a practical authority plan.
Reviews
Review volume, quality, freshness, and platform coverage can shape trust.
Directories
Trusted local, industry, and category directories can become source signals.
Mentions
Independent mentions help answer engines corroborate claims about the business.
Video
YouTube and other video sources can influence AI-shaped discovery in many markets.
Proof pages
Case studies, customer stories, comparison pages, and outcomes make claims easier to support.
Consistency
Names, addresses, service descriptions, and categories should match across the web.
Step 5
Measure each answer engine separately.
There is no single AI search engine. A business can appear in Gemini and be absent from Perplexity. It can be mentioned by ChatGPT for one prompt and ignored for another. The measurement layer should track the exact prompt, the exact answer, the competitors named, the citations used, and the movement over time.
Public sources
Credibility starts with what the platforms actually publish.
AI SEO is moving quickly, so Alchemetics separates official platform documentation from vendor studies and marketing claims. The hard percentages in this market should be treated as directional until your own prompt tracking proves them in your category.
OpenAI crawler documentation
OpenAI publishes crawler information for GPTBot, OAI-SearchBot, and ChatGPT-User at developers.openai.com.
Anthropic crawler documentation
Anthropic documents ClaudeBot and related crawler behavior for site owners.
Perplexity crawler documentation
Perplexity publishes crawler and user-agent information for site owners.
Bing IndexNow
IndexNow gives site owners a way to notify participating search engines about changed URLs.
Google Search Central
Google’s SEO fundamentals still matter for crawlability, indexing, structured data, and page quality.
Market testing
The final ground truth is prompt tracking in the client’s actual category, not a generic industry score.
FAQ
Questions about AI search readiness.
What is AI Search Readiness?
It is the technical, content, entity, trust, and measurement foundation that helps answer engines read, understand, cite, and recommend a business.
Can a website rank in Google but fail in AI search?
Yes. Ranking is only one signal. A site can rank but still be hard for AI systems to extract, verify, or cite.
Should I block AI training crawlers?
That is a business decision. The key is not to accidentally block search or retrieval crawlers that may support visibility.
Do AI engines all use the same sources?
No. Different engines use different retrieval systems, indexes, source preferences, and citation behavior.
Is llms.txt required?
No. It is a low-cost forward-looking file, but it should not be sold as a current ranking factor or citation guarantee.
What is the first audit Alchemetics runs?
We start with access and readability: robots rules, status codes, redirects, raw HTML, schema, indexation, and whether the page answers real buyer questions.
Get the report
Find out whether AI can read, trust, and recommend your business.
The AI Visibility Guru report shows the real prompts, answers, competitors, sources, and first fixes that matter most.