Which buyer questions should you track first?
Start with the questions a buyer would ask before they know your company, not with searches containing your brand name. Group the prompt set by problem, solution category, comparison, use case, and purchase decision. Write each question as a real person might ask it in ChatGPT, Perplexity, Gemini, Claude, Grok, Google AI Overviews, or Google AI Mode.
Keep the first set small enough to inspect manually. Include questions where an answer could recommend a provider, compare alternatives, explain a process, or cite evidence. Add the important variations in terminology, such as the category name, the job title of the buyer, and the language used by sales calls.
Record the intended audience and buying stage beside every prompt. That context prevents a common launch mistake: treating a branded answer as proof that the new domain is visible for category questions. A useful baseline separates branded prompts from unbranded discovery prompts and separates informational questions from commercial ones. The resulting list becomes the test set for every launch decision, rather than an unfocused collection of queries.
For more context, read AI Search Technical SEO Checklist: What to Fix First.
How do you establish a baseline before launch?
Capture the current answers before the new domain goes live, then save the exact wording, cited sources, mentioned competitors, and answer date for each prompt. A pre-launch baseline is possible even when the domain has no history because the baseline measures the market's current answer landscape, not your site's traffic.
For each engine, note whether your company is mentioned, whether a source is cited, and whether the cited page belongs to a competitor, an independent publisher, or an official source. Also record where your existing information lives, if it is already published elsewhere. This distinction matters for a rebrand or domain change, but it also matters for a genuinely new company whose evidence is scattered across public sources.
Do not average all engines into one launch score. ChatGPT, Perplexity, Gemini, Claude, Grok, Google AI Overviews, and Google AI Mode can return different sources and wording. Preserve the raw answers so a later change can be traced to a specific prompt and engine. A baseline is useful only when the same prompt, wording, and observation rules are used again.
For more context, read AI Visibility for Product Launches: What to Do First.
What must you verify before engines can reach the domain?
Confirm that search crawlers can fetch the pages that support your main buyer questions before judging AI visibility. Check robots.txt, noindex directives, canonical URLs, HTTP status codes, XML sitemaps, internal links, and whether important content appears in the rendered page rather than only after an interaction.
Use a small set of representative URLs: the homepage, a category page, a problem page, a comparison page, and one evidence page. Check that each resolves to the intended host, uses one canonical version, and links to related pages with descriptive anchor text. Submit the sitemap and inspect indexing and crawl signals in Google Search Console, while remembering that search indexing does not guarantee inclusion in an AI answer.
The launch order matters. Fix access and duplication problems before rewriting copy, because a strong page cannot help an engine that cannot reliably retrieve it. Keep a record of the checks and their dates. When an answer later cites an old URL, a staging host, or a thin page, the record helps distinguish a technical discovery problem from a content problem.
Which pages should receive attention on launch day?
Prioritize pages that can directly answer high-value unbranded questions and give an engine a clear source to cite. Choose one primary page for each important intent, then make sure that page states the answer, defines the audience, explains the relevant trade-offs, and links to supporting evidence.
A homepage should establish what the company does and who it serves, but it should not carry every category question. A category or solution page usually deserves priority for broad discovery prompts. A comparison or decision page can support commercial questions, while a proof page should substantiate claims rather than repeat positioning. Avoid publishing several near-identical pages for one prompt, because competing internal answers make citation selection less clear.
Use the baseline to rank pages by opportunity. A page is a stronger first candidate when an engine already answers the question, cites a competitor, and your new domain has a credible page that could address the same need. A page with no clear audience or evidence is not automatically the next fix. The launch checklist should favor answerable demand over the number of URLs published.
How do you prevent a new domain from looking unrelated?
Make the relationship between the company, domain, products, people, and evidence unambiguous across the public web. A new domain can be technically accessible yet appear unrelated if its company name, descriptions, contact details, author information, and external references disagree.
Use the same factual naming wherever the company is represented. State what the company does, which buyers it serves, the problems it addresses, and how product or service names relate to the company. Link relevant profiles and existing references to the new domain where appropriate, and make the relationship clear rather than relying on an engine to infer it from a logo or design.
Apply a decision rule for rebrands and new companies. If the company existed under another domain, preserve useful redirects and explain the relationship on key pages. If the company is new, do not manufacture history or imply third-party validation that does not exist. Instead, publish clear first-party explanations and evidence that can be checked. Test branded and unbranded prompts separately. A branded answer can confirm recognition while an unbranded answer still fails to connect the company with the category.
What should you change when an engine cites a competitor?
First identify why the competitor was selected, then fix the missing answer rather than adding the competitor's name repeatedly. Compare the cited page with your intended page for directness, evidence, specificity, freshness, and fit with the question. The useful question is not only, “Why was our brand absent?” but also, “What did the cited source make easy to verify?”
If the competitor explains a definition better, improve your explanation. If it provides a comparison, address the decision criteria plainly. If it is cited for independent evidence, add original, supportable evidence or point readers to a source that can verify the claim. If your page covers the subject but buries the answer, move the relevant explanation closer to the top and connect it to the supporting page.
Do not make every change at once. Assign one prompt to one proposed fix, publish the smallest useful change, and preserve the previous version for comparison. Then rerun the same prompt across the affected engines. This process exposes a failure mode that launch teams often miss: a page can gain mentions while remaining a poor citation because its claims are difficult to verify.
How should you compare the seven engines at launch?
Compare the seven engines as separate answer environments, not as one blended AI channel. The set includes ChatGPT, Perplexity, Gemini, Claude, Grok, Google AI Overviews, and Google AI Mode. Google AI Overviews and Google AI Mode should be recorded separately because their answer experiences and citation displays are not interchangeable.
Use the same prompt, location assumptions, language, and observation date wherever the testing method permits. Record mention status, citation status, cited URL, competitor presence, and the type of source used. A brand mention without a citation is a different outcome from a cited recommendation, and a citation to a category page is different from a citation to a specific proof page.
Avoid treating absence in one engine as proof of a site-wide failure. Engines may retrieve different sources, use different freshness signals, or answer the same wording differently. Cituna tracks whether these seven engines mention and cite a brand for buyer questions every day, and shows which competitors and pages they cite instead. Cituna does not track Microsoft Copilot, so Copilot results require a separate testing process if they matter to your audience.
When should you choose the next launch fix?
Choose the next fix when a repeated, commercially important prompt shows a clear gap that your domain can address. Give priority to a prompt that appears across several relevant engines, leads to a meaningful buyer decision, and has a credible page or evidence gap you can correct. Do not prioritize a one-off wording change from a single volatile answer over a recurring absence from important questions.
After publishing a fix, rerun the same prompt set and inspect the cited pages, not just the presence of your company name. Join answer observations with Google Search Console data to see whether the page is receiving search impressions and whether its query coverage matches the intended buyer questions. Search visibility and AI citation are related signals, but neither substitutes for the other.
Cituna joins daily observations from all seven tracked engines to Google Search Console data and provides SEO, AEO, and GEO fixes. Every engine is included on every plan without a per-engine add-on fee. The entry plan is $39 per month and covers 10 tracked prompts, with Search Console and an MCP server included; the API is available on Max. Choose a tool only after deciding which prompts and decisions the launch process must measure.
Related reading
- How To Compare Ai Visibility Measurement Before Buying
- AI Visibility Launch Checklist for New Markets
Sources consulted
- Google Search Central (developers.google.com)
- OpenAI (platform.openai.com)
- Perplexity (docs.perplexity.ai)
- Anthropic (anthropic.com)
Drafted with AI assistance from our own research and Search Console data, and reviewed by Rahul A before publishing. Rules and prices change; check the linked official source before you act.