LLM SEO / GEO operating instructions
In short: This prompt helps with: lLM SEO / GEO operating instructions. Nothing to fill in: copy it as it is.
Want more like this?
Shared publicly by Alex Havryleshko (@alexhavryleshko) on X. Text is as posted. X profile · Original post · Report or remove
LLM SEO / GEO operating instructions
Last verified: 2026-10-04
Goal: get pages retrieved and cited in ChatGPT search, Perplexity, Claude, and Google AI Overviews / AI Mode. This does not replace SEO. Retrieval still depends on being crawlable and, for Google, indexed with a snippet.
Finance is YMYL. The retrieval steps do not change. The citation bar does.
What is verified
Google: a page must be indexed and eligible for a snippet. There are no extra technical requirements for AI Overviews or AI Mode. Structured data is not required for generative AI search, and there is no special schema to add. llms.txt and other special AI markup are not used by Google Search. Source: Google Search Central, AI features and AI optimization guide.
OpenAI: OAI-SearchBot builds the index used in ChatGPT search. Blocking it removes the site from ChatGPT search results (navigational links can still appear). GPTBot is training only. The two robots.txt rules are independent. Source: OpenAI bots docs.
GEO paper (Aggarwal et al., KDD 2024, arXiv:2311.09735): on already-retrieved sources, adding citations, quotations, and statistics raised position-adjusted word count about 30–40%, and up to ~40% visibility overall. On Perplexity the same methods reached up to 37%. Gains were largest for lower-ranked sources (Cite Sources +115% at rank 5, negative at rank 1). Keyword stuffing did not help. This was a controlled rewrite test, not a ranking study.
Unverified or third-party only (do not treat as fact): Wikipedia/Reddit citation-share figures, an exact 40–60 word answer length, “100+ articles = 3.4× citations,” and day-to-day Jaccard noise bounds. Use them as hypotheses.
Order of work
Access
Google eligibility
Quotable pages
Off-site corroboration
Measurement
Do not start with schema, llms.txt, or content rewrites if crawlers cannot reach the page.
1. Access
Allow search crawlers in robots.txt. Do not block them at the CDN or WAF.
User-agent: OAI-SearchBotAllow: /
User-agent: PerplexityBotAllow: /
User-agent: Claude-SearchBotAllow: /
User-agent: GooglebotAllow: /
User-agent: bingbotAllow: /
Training bots (GPTBot, ClaudeBot, CCBot, Google-Extended) are a separate policy. Blocking them does not remove you from search.
Confirm in server logs that OAI-SearchBot and PerplexityBot actually fetch HTML. If they never appear, copy changes will not help.
Serve the answer as HTML text. Do not hide the main content behind client-only render. Fees, APRs, and eligibility tables must be in the HTML, not only in an image or a logged-in app.
2. Google eligibility
Index the URL. Check Search Console.
Do not nosnippet pages you want in AI Overviews or AI Mode.
Internal-link important pages.
Keep Merchant Center and Business Profile accurate if you sell or have a location.
Schema is optional for AI features. If used, it must match visible text. Use it for rich results, not as an AI ranking trick.
Do not add llms.txt expecting a Google effect. Some non-Google agents may read it; that is optional and unproven for citations.
3. Page pattern
One page, one question. Answer it in the first paragraph. Then evidence. Then limits.
# {Question as the H1}
{Direct answer in 2–4 sentences. Name the legal entity. Include one number, its as-of date, jurisdiction, and source.}
## {Sub-question}
{Answer in the first sentence. Then the stat or quote. Then the source link.}
Rules:
Specific facts only: fee, APR, limit, date, version, geography. Adjectives are not citable.
Cite the primary source next to the claim. For finance that source is the fee schedule, prospectus, rate card, or regulator page, not a blog.
Put a visible author name and a visible updated date on the page.
Keep the legal name, product name, pricing, and definition identical to filings, the site, and public profiles.
Refresh when the fact changes. An undated rate is worse than no rate.
Do not keyword-stuff. The GEO paper found it did not help.
Do not add hidden instructions for models. Providers treat that as spam.
llms.txt and FAQ schema are optional hygiene. They are not the lever.
4. Fintech
State jurisdiction, license or registration, and the regulator on every product page.
Put the as-of date on every rate, fee, limit, and yield. Models drop undated figures.
Match comparison tables to the source exactly. Do not round a fee into a claim the table does not support.
Do not write "best," "guaranteed," or personalized advice. Eligibility and risk belong in the same passage as the number.
Cite the regulator, central bank, filing, or your own fee schedule. Third-party roundups are weak sources for a rate.
Do not cite a number you cannot show in HTML on that page.
5. Off-site
Models cite pages they already retrieve. Third-party pages often outrank your own.
Get mentioned on sites already cited for the query: regulator pages, filings, docs, comparison lists, review sites.
For Perplexity-style answers, useful community threads matter more than a new blog post.
For Google AI features, ordinary rankings still gate eligibility. A page that is not indexed cannot be a supporting link.
Do not buy mentions. Google warns against inauthentic mentions.
6. Measurement
Pick 20–50 real buyer prompts. Include fee, eligibility, license, and "X vs Y" prompts. Run each in ChatGPT, Perplexity, and Google AI Mode. Record: cited URL, brand named or not, whether the fact, date, and jurisdiction are correct.
Repeat the same prompts. Citation sets move between runs. Ignore small swings.
Useful proxies: Search Console impressions, Bing AI citation reports if available, log hits from search bots.
Success is a correct citation on a prompt you care about, not a single screenshot. A citation with the wrong rate is a failure.
Do not
Claim a special AI schema or llms.txt ranking boost on Google.
Block OAI-SearchBot and expect ChatGPT citations.
Treat the GEO paper’s 40% as a traffic guarantee. It measured share of an answer among sources already in context.
Publish a thin page per keyword variant. Answer the fan-out on the page that already ranks.
Publish a rate, yield, or "best product" claim without jurisdiction, as-of date, and a primary source.
Fill in the blanks (optional)
Type your details and the prompt above updates. Copy uses what you've filled in; blanks stay as [PLACEHOLDERS].
Fill in the [PLACEHOLDERS] before sending. Never paste real customer data into an AI tool you don't control.
Got a prompt that works for you? Add your prompt. It takes a minute.
Sourced from a public post
Some links are affiliate or sponsored: we may earn a commission at no cost to you. Sponsor this spot →
Related prompts
MarketingChatGPTPartnership announcement on LinkedInAnnouncing a new partnership publicly.View prompt →MarketingChatGPTFacebook campaign and ad copyPlanning a month of paid social for an offer.View prompt →MarketingChatGPTTurn a campaign plan into a marketing briefHanding a campaign to a designer, agency or teammate.View prompt →MarketingChatGPTLinkedIn hiring postYou have a full job description and want a short post to attract applicants.View prompt →
Browse: Marketing, LinkedIn & content
Something wrong with this prompt? Contact us or submit a better one.