# Win the prompts that matter, with a team that ships the pages. > See which prompts you win on ten AI engines and which pages get cited, find the topics rivals cover, and ship citable rewrites to WordPress or Webflow. [HTML version](https://www.citeme.io/solutions/seo-content) --- A Google position does not tell you whether an AI answer cites you. Our team reads the answers engine by engine, finds the prompts you lose and what competitors cover, then rewrites and publishes the fix on your CMS. You approve every change on a shared board. - Prompts won and lost, page by page - Rewrites and new pages, published for you - You approve every change on a shared board ## Your page ranks. The AI answer quotes someone else. Many sites with strong SEO are invisible in AI answers. An engine retrieves a passage, not a URL, so an article whose key claim is spread across four paragraphs can lose to a plainer one that states it in a single block. And your Google position does not say whether you are among the one to three sources cited. ## Prompts, pages, passages, measured in AI answers. ### Prompts won and lost Prompts are generated from your sitemap in three families: brand, comparative, problem or solution. Each answer records whether you are named, your position, your sentiment and the competitors cited alongside. ### Citations, tied to pages Citeme records which answers cite which of your pages. Lose first place on a query and an alert names the competitor that displaced you, by email and on the dashboard. ### The next page to write Semantic gap analysis names the subjects, concepts and content types cited competitors cover and you do not: the actual reason they are cited and you are absent. ### Passages engines can lift An engine retrieves a passage, not a URL, so every block must answer on its own. Citeme's working target: around five verifiable facts per section of 200 to 300 words. ### The technical layer, patched A light audit scans robots.txt access for AI crawlers, JSON-LD, metadata and heading hierarchy without calling any model. Nova then patches titles, meta, schema.org, alt text and hreflang page by page. ### Bot reads, page by page Which pages AI bots read, how often and by which provider. A page never read has an access or linking problem; one crawled constantly but never cited points at what it says. ## Which prompts you lose, and what to ship next. ### Read every answer, per prompt Per keyword, open the raw answer from each model: your exact position, the competitors cited alongside you, your sentiment and the excerpts showing how the model describes you. Each keyword is asked zero-shot, chain-of-thought and few-shot. ### Rank what to write next Nova reads the audit, spots the gaps against competitors and ranks the next moves: an existing page that lost a citation to rewrite, a prompt where a rival is cited that needs a new page, a weak title or meta to fix. ### Rewrite, then publish Nova rewrites an existing page to make it citable (original data, FAQ structure, clarity) or writes a new blog post, landing page or FAQ. Once you approve, it ships to WordPress or Webflow with JSON-LD, and Search Console submits the URL for reindexing. ### Tie each visit to its page A visit from an AI answer is matched to the prompt it most likely came through, via the answers citing its landing page. A page answering one precise question attributes cleanly; a generic page cited for a dozen prompts only gives a guess. ## FAQ ### How is GEO different from SEO? SEO optimizes for ranking in a list of blue links. GEO optimizes for being the source an AI cites or paraphrases inside a single generated answer: different signals, different content structure, a different scoreboard. The two are linked: generative engines rely on traditional indexes to find their sources, so a site Google cannot crawl stays invisible to ChatGPT. ### Do we need to rewrite our best SEO pages? Often they need restructuring rather than replacing. The substance is already there, but the key claim is spread across several paragraphs instead of being stated completely in one. That is an editing job rather than a writing job, and the cheapest visibility work available to a site that already ranks well. ### Why is one of our pages never cited? Check access first. Pages behind a login, marked noindex or absent from your sitemap are never indexed, so they never feed prompt generation; behind Cloudflare or a strict WAF, allow the CiteMeBot/1.0 user agent. Then read the per-page breakdown: a page no AI bot opens has an access or linking problem, one read but never cited points at what it says. ### Can it publish straight to our CMS? Yes. WordPress connects over the REST API with an application password, version 5.6 or later, no mandatory plugin; Webflow over OAuth, to manage a blog post, optimise a page or update the whole site. Drafts or live, your call, and nothing Nova prepares is published before someone approves it. Shopify connects the same way, and Nova publishes there too. ### What about llms.txt? It is a plain text file at your domain root telling AI crawlers what your site is about and which pages matter: roughly twenty minutes of work for a site of normal size, and Citeme has a free generator. Adoption is still partial in 2026, so treat it as a cheap complementary signal. On a site whose AI crawlers are blocked, it produces nothing. ### How do we know a rewrite worked? Two signals. The live feed shows within hours whether the crawlers came back to read the page. Citations move with the next audits: the score blends runs, so a real improvement shows over two or three audits, and around five points of movement is ordinary model fluctuation. ## Related pages - [Citations](https://www.citeme.io/features/citations): See who gets cited, where and how - [Nova](https://www.citeme.io/features/ai-agents): The AI assistant that fixes your GEO - [In-house teams](https://www.citeme.io/solutions/in-house): Run GEO yourself, from one platform - [Pricing](https://www.citeme.io/pricing): Plans and the done-for-you engagement