How do I map every old and new product URL?
Start by mapping every old product URL to its intended destination before changing pages, because AI visibility can fall when a consolidation removes the page that previously supplied a clear answer. Export the old URLs, current URLs, status codes, titles, headings, canonical tags, internal links, XML sitemap entries and important structured data. Add the product, audience, use case and buyer question each URL served.
Mark each old page as keep, merge, redirect, retire or investigate. A redirect target should answer the same underlying buyer need, not merely share a product category. A page that covered implementation may need a different destination from a page that covered features or comparisons. Preserve those distinctions in the map even if several pages become one URL.
Check whether external references, partner links or saved campaign links still point to pages being removed. Also check whether the proposed destination already serves a different intent. The most useful consolidation map is therefore an intent map, not just a spreadsheet of URLs. Teams with several related offers can use Manage AI Visibility Across Product Lines to examine how product relationships affect discovery.
Choose one canonical page for each buyer intent
Choose one canonical page for each buyer intent, even when several products share similar features or terminology. A consolidated page should make clear which problem it solves, who it is for, what it includes and how it differs from nearby offers. Without those distinctions, an AI assistant may cite the brand while naming the wrong product, or omit the brand because the answer appears ambiguous.
Compare the old pages by evidence rather than by preference. Check organic clicks, impressions, conversions, assisted conversions, backlinks, meaningful internal links and the quality of the page's answers to buyer questions. Search Console can indicate which URL receives demand, but it does not by itself prove that the URL is the best destination for every intent.
Write a short decision record for each merge. State which page survives, which intents it inherits, which claims need combining and which claims must be removed. Keep separate destinations when audiences, use cases, eligibility, deployment models or product capabilities genuinely differ. If the surviving page becomes too broad to answer a question directly, consolidation may reduce visibility instead of improving it.
Redirect removed URLs to the closest useful destination
Redirect each removed product URL to the closest useful replacement, and never send every retired page to the homepage by default. A redirect preserves a route to the replacement, but it cannot transfer meaning to a destination that does not answer the original question. Check that old URLs return the intended permanent redirect, resolve to the preferred hostname and protocol, and reach the final page without a redirect chain.
Test representative URLs from every retired product, feature, comparison and documentation pattern. Include URLs with parameters, trailing slash variations, uppercase characters and legacy folder structures. Check that the destination loads for ordinary visitors and that its title, main heading and visible copy match the redirect decision.
Review links from navigation, related-product modules, support content, sales pages and downloadable resources. Internal links that still point to retired URLs create avoidable ambiguity and can leave crawlers following a route that no longer represents the site structure. If consolidation also changes the market position or promise, use Manage AI Visibility After Repositioning as a separate check rather than assuming the redirect work covers it.
Align canonical signals and product structured data
Align canonical tags, visible copy, structured data and internal links so every signal identifies the same surviving product page. A canonical tag is only one signal, so check whether it agrees with the page linked from navigation, the URL in the sitemap, the page's redirects and the product information shown to users.
Review Product, Offer, Organization, BreadcrumbList and FAQ markup where those types accurately describe the page. Remove markup inherited from a retired product if the claims no longer apply. Check names, descriptions, identifiers, availability, pricing and brand relationships against the visible page. Structured data should clarify the page, not preserve an old product identity after a merge.
Make the surviving page explicit about former names or related products only when that context helps buyers and remains accurate. Do not keep multiple near-identical pages live solely to preserve old wording. Google may treat them as competing signals, while AI assistants can encounter different descriptions depending on which URL they retrieve. A clean consolidation gives each product question one strong destination and one consistent set of facts.
Remove stale pages, links and discovery paths
Remove stale pages and discovery paths after redirects are working, because an old page that remains crawlable can continue competing with the consolidated destination. Check the XML sitemap, HTML sitemap, category pages, search results, related-product blocks, navigation menus, feeds and CMS-generated archives. A retired URL should not appear as an active product in any of those places.
Check the page source and rendered content for old product names, discontinued feature claims, conflicting identifiers and links to other retired URLs. Review PDFs, help articles and comparison pages that may still describe the old product as current. The aim is not to erase historical references that buyers need, but to prevent outdated pages from presenting themselves as valid alternatives.
Use server logs or crawler data to identify repeated requests for removed URLs. A persistent request can come from a valuable external reference, a broken internal link or an automated system. Investigate the source before blocking it. Robots rules can stop crawling, but they do not replace a useful redirect or remove a page from every system that has already recorded it. The final check is whether a visitor or crawler can still discover two apparently current answers to one product question.
Create a before-and-after prompt baseline
Create a before-and-after prompt baseline using the same buyer questions, engines and scoring rules, because visibility changes are easy to misread after a large URL change. Build prompts around category discovery, product recommendations, comparisons, alternatives, use cases, implementation and former product names. Record the exact wording, location, date, engine and any relevant personalization controls.
Run the baseline across ChatGPT, Perplexity, Gemini, Claude, Grok, Google AI Overviews and Google AI Mode. Record whether the brand appears, the product named, the cited or linked page, the position of the mention and the competitors shown instead. Note whether an answer is correct but unsupported, supported by an old URL or supported by a page that no longer matches the recommendation.
Repeat prompts after redirects, indexing and content changes have had time to be observed. Compare like with like rather than treating one answer as proof of a trend. Cituna asks all seven engines the questions a brand's buyers ask every day and records names, citations, positions, competitors and replacement pages. That makes the consolidation test about changed answers and sources, not only about rankings in traditional search.
Fix the gap revealed by each changed answer
Fix the specific gap revealed by each changed answer before rewriting the entire product page. If an engine names the brand but selects an old URL, repair redirects, internal links and canonical signals. If it cites a competitor for a question the consolidated page should answer, strengthen the missing explanation, comparison, proof or use-case detail. If it recommends the wrong product, clarify the boundary between products and add direct language for the intended buyer.
Check the page against the prompt that exposed the gap. Add an answer only when it is accurate, supported by the product, and useful to a human reader. A broad page can need clearer sections rather than more words. A missing technical fact may need documentation or a specification table. A weak comparison may need explicit differences, not repeated brand claims.
Manual review works when the prompt set is small and the team can inspect every answer. A platform is more useful when the team needs daily scans, tracked prompts and a path from a detected gap to a proposed fix. Cituna generates schema, FAQ markup, llms.txt and page changes for detected gaps, and its AutoSEO can write and publish articles from those gaps and Search Console demand. Approval should remain with the person responsible for product accuracy.
Set a post-consolidation review rule
Set a post-consolidation review rule that separates a technical failure from a content failure and a product decision. Review technical signals first, including status codes, canonical URLs, sitemap entries, crawl access and internal links. Review answer accuracy next, including the named product, cited page and competitor appearing instead. Only then decide whether the consolidated page needs new content or whether the consolidation itself should be reversed.
Use Search Console to connect page changes with clicks and queries, while engine checks show whether assistants name and cite the page. A click increase without better answer accuracy may not solve recommendation problems. Better mentions without visits may indicate that the cited page lacks a strong next action or serves the wrong intent.
Keep the old prompt set unchanged for trend comparison, then add new prompts only when the product, audience or terminology changes. Recheck after substantial redirects, page edits, indexing changes or competitor shifts. If the surviving page repeatedly answers several unrelated intents poorly, split it deliberately rather than adding more sections indefinitely. The decision rule is simple: keep one page when it gives one coherent answer, and separate pages when one URL creates competing answers.
Related reading
- AI Search Technical SEO Checklist and Fix Order
- AI Visibility Tool Costs: Pricing Models and Budget Rules
Sources consulted
- Google Search Central (developers.google.com)
- Google Search Console Help (support.google.com)
- OpenAI Platform (platform.openai.com)
- Perplexity API Documentation (docs.perplexity.ai)
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.