Everything your audit checks.
Six capabilities, one report. Every audit measures all of them and surfaces the fixes ranked by expected score lift — capability by capability, here is what RankLens looks at.
Live LLM probing
AI engines do not index pages — they answer questions. RankLens runs live category prompts against ChatGPT, Claude, Perplexity, and Gemini and records exactly what each engine says about your category: verbatim responses, which stores it cites, how it characterizes your brand, and the sentiment score attached to each mention.
Most SEO tools tell you where you rank on Google. They do not tell you whether ChatGPT names you when a buyer asks for the best running shoes for flat feet. Live probing does.
- Verbatim AI responses — the exact text each engine returned, stored per audit so you can diff week over week.
- Citation records — which domains each engine linked or named, and how often yours appeared.
- Sentiment scoring — positive, neutral, or negative characterization of your brand in AI-generated answers.
- Multi-engine coverage — ChatGPT, Claude, Perplexity, and Gemini queried in parallel, so you see where your visibility is strong and where it drops.
Variant-level depth
Most audit tools check your homepage and a handful of top-level pages. RankLens crawls every product detail page — including color, size, and material variants — because that is the surface AI engines extract structured data from when a buyer asks a specific product question.
Schema markup at the variant level is how AI models read your catalog. A parent PDP with valid JSON-LD but missing variant-level price, availability, and SKU fields is a partial schema — and a partial schema produces partial AI answers.
- JSON-LD completeness — required and recommended fields present at the variant level, not just the parent product.
- Correctness — field types, value formats, and nested structures validated against the schema.org Product spec.
- AI extraction confidence — a per-field score reflecting how reliably each AI engine can parse the value.
- Structured data depth — coverage across your full catalog, not a sampled subset.
Prescriptive playbook
An audit that surfaces issues without telling you what to do with them is a list, not a plan. The RankLens playbook turns every finding into a ranked fix with three fields: what to change, where to change it, and the expected score lift from making that change.
Fixes are ordered by impact, not by severity alone — a low-severity fix that applies to 200 PDPs ranks above a high-severity fix that applies to 1. The ranking accounts for catalog breadth, AI extraction confidence delta, and the effort cost of applying each fix class.
- Fix description — a plain-language statement of what the issue is and what the corrected state looks like.
- Copy-paste implementation — the exact JSON-LD snippet, meta tag value, or heading text to add or replace.
- Expected score lift — the estimated change in AI extraction confidence if the fix is applied correctly.
- Affected PDP count — how many pages carry the issue, so you can prioritize by catalog impact.
Shopify one-click apply (Coming soon)
Generating a fix list is one step. Applying it to live PDPs without touching code is another. The RankLens Shopify embedded app connects to your store once and surfaces every playbook fix inside the Shopify admin — review the change, confirm, apply. No theme edits, no developer handoff, no CSV exports.
The embedded app is built on Shopify App Bridge so it lives in your existing admin workflow. Authentication is handled once at connection; subsequent audit runs and fix applications do not require re-auth.
- Schema injection — adds or replaces JSON-LD blocks on PDPs and variant pages without touching your theme liquid.
- Meta tag updates — title, description, and structured meta fields updated at the page level.
- Batch application — apply a fix class across all affected PDPs in one action rather than page by page.
- Rollback — every applied fix is versioned; revert to the previous state from the audit history view.
Continuous monitoring
AI engines update their answers weekly. A store that appears in ChatGPT's recommendations today may be absent next week if a competitor improves their schema, if the engine updates its training context, or if your product catalog drifts from the structured data it was audited against.
RankLens re-probes your tracked prompts every week and surfaces a diff: which citations appeared, which disappeared, and how your sentiment score moved. Monitoring catches drift before it becomes a traffic event.
- Weekly re-probes — the same category prompts run against ChatGPT, Claude, Perplexity, and Gemini on a rolling weekly cadence.
- Diff alerts — email notification when your citation share changes by a threshold you set, so you are not logging in to check manually.
- Sparkline trends — per-prompt, per-engine visibility scores plotted over time in your audits view, so directional movement is visible at a glance.
- Prompt library — add, edit, or remove tracked prompts as your catalog or category focus changes.
Sources analysis
Knowing whether AI names your store is the first question. Knowing which sources AI cites to answer your category queries — and what share of those citations go to competitors — is the second.
Sources analysis maps the citation landscape for your category. For each tracked prompt, RankLens records every domain an AI engine cites, the frequency of citation, and the share each domain holds across the full response set. Your store's citation share is plotted against the top-cited competitors, so you can see exactly where the gap is and how it moves over time.
- Citation share by domain — your store's share of AI citations for each tracked prompt, expressed as a percentage of total citations recorded in that probe session.
- Competitor citation map — which competitor domains AI engines cite most often in your category, and how their share has changed week over week.
- Source type breakdown — citations to product pages, review sites, editorial content, and brand homepages, so you know which content types AI engines draw from in your category.
- Gap prioritization — the playbook cross-references sources data to surface fixes that are likely to improve citation share, not just extraction confidence.
Agencies
Agencies managing 5+ stores get the most value from the prescriptive playbook — ranked fixes per audit, copy-paste-ready, and exportable as a white-label PDF for the client.
Jump to the playbook →SEO managers
SEO managers measuring AI-search visibility start with live LLM probing — verbatim engine responses, citation records, and sentiment per audit — then move to sources analysis to see citation share.
Jump to live LLM probing →Shopify stores
Shopify stores apply fixes from inside the admin via the Shopify one-click apply workflow — schema injection and meta updates with no theme edits and a versioned rollback.
Jump to Shopify one-click apply →