{ "schema_version": 2, "measurement_set": "f27245a8e3c26349", "settings": { "perplexity": { "provider": "openrouter", "model": "perplexity/sonar-pro", "search": "native", "request_version": 2, "question_models": {}, "returned_models": [] }, "openai": { "provider": "openrouter", "model": "openai/gpt-5.2:online", "search": "native", "request_version": 2, "question_models": { "measure-problem": "openai/gpt-5.2", "measure-category": "openai/gpt-5.2", "measure-decision": "openai/gpt-5.2" }, "returned_models": [ "openai/gpt-5.2" ] }, "anthropic": { "provider": "openrouter", "model": "anthropic/claude-sonnet-5:online", "search": "native", "request_version": 2, "question_models": {}, "returned_models": [] }, "gemini": { "provider": "openrouter", "model": "google/gemini-2.5-flash:online", "search": "native", "request_version": 2, "question_models": {}, "returned_models": [] } }, "domain": "skilltrade.marketing", "competitors": [], "arm": null, "run_at": "2026-09-10T15:11:22.676852+00:00", "prompt_set": "033e048118d3", "prompt_count": 3, "_prompts": [ "How can a B2B marketing team tell whether AI answers name its brand?", "What kinds of tools can a B2B marketing team use to track citations in AI answers?", "Which free, open-source tools can a B2B marketing team use to track whether AI answers cite its website?" ], "questions": [ { "id": "measure-problem", "stage": "problem", "text": "How can a B2B marketing team tell whether AI answers name its brand?" }, { "id": "measure-category", "stage": "category", "text": "What kinds of tools can a B2B marketing team use to track citations in AI answers?" }, { "id": "measure-decision", "stage": "decision", "text": "Which free, open-source tools can a B2B marketing team use to track whether AI answers cite its website?" } ], "brand_aliases": { "skilltrade.marketing": [ "Skill Trade", "SkillTrade" ] }, "audience": {}, "engines": { "perplexity": { "prompts_run": 0, "prompts_attempted": 3, "status": "incomplete", "cited": 0, "citation_share": null, "observed_share": null, "competitor_share": {}, "stages": { "problem": { "planned": 1, "completed": 0, "brands": { "skilltrade.marketing": { "counts": { "mentioned": 0, "owned_cited": 0, "uncited_mention": 0, "third_party_candidates": 0 }, "rates": { "mentioned": null, "owned_cited": null, "uncited_mention": null, "third_party_candidates": null } } } }, "category": { "planned": 1, "completed": 0, "brands": { "skilltrade.marketing": { "counts": { "mentioned": 0, "owned_cited": 0, "uncited_mention": 0, "third_party_candidates": 0 }, "rates": { "mentioned": null, "owned_cited": null, "uncited_mention": null, "third_party_candidates": null } } } }, "decision": { "planned": 1, "completed": 0, "brands": { "skilltrade.marketing": { "counts": { "mentioned": 0, "owned_cited": 0, "uncited_mention": 0, "third_party_candidates": 0 }, "rates": { "mentioned": null, "owned_cited": null, "uncited_mention": null, "third_party_candidates": null } } } } } }, "openai": { "prompts_run": 3, "prompts_attempted": 3, "status": "complete", "cited": 1, "citation_share": 0.3333333333333333, "observed_share": 0.3333333333333333, "competitor_share": {}, "stages": { "problem": { "planned": 1, "completed": 1, "brands": { "skilltrade.marketing": { "counts": { "mentioned": 0, "owned_cited": 0, "uncited_mention": 0, "third_party_candidates": 0 }, "rates": { "mentioned": 0.0, "owned_cited": 0.0, "uncited_mention": 0.0, "third_party_candidates": 0.0 } } } }, "category": { "planned": 1, "completed": 1, "brands": { "skilltrade.marketing": { "counts": { "mentioned": 0, "owned_cited": 0, "uncited_mention": 0, "third_party_candidates": 0 }, "rates": { "mentioned": 0.0, "owned_cited": 0.0, "uncited_mention": 0.0, "third_party_candidates": 0.0 } } } }, "decision": { "planned": 1, "completed": 1, "brands": { "skilltrade.marketing": { "counts": { "mentioned": 1, "owned_cited": 1, "uncited_mention": 0, "third_party_candidates": 0 }, "rates": { "mentioned": 1.0, "owned_cited": 1.0, "uncited_mention": 0.0, "third_party_candidates": 0.0 } } } } } }, "anthropic": { "prompts_run": 0, "prompts_attempted": 3, "status": "incomplete", "cited": 0, "citation_share": null, "observed_share": null, "competitor_share": {}, "stages": { "problem": { "planned": 1, "completed": 0, "brands": { "skilltrade.marketing": { "counts": { "mentioned": 0, "owned_cited": 0, "uncited_mention": 0, "third_party_candidates": 0 }, "rates": { "mentioned": null, "owned_cited": null, "uncited_mention": null, "third_party_candidates": null } } } }, "category": { "planned": 1, "completed": 0, "brands": { "skilltrade.marketing": { "counts": { "mentioned": 0, "owned_cited": 0, "uncited_mention": 0, "third_party_candidates": 0 }, "rates": { "mentioned": null, "owned_cited": null, "uncited_mention": null, "third_party_candidates": null } } } }, "decision": { "planned": 1, "completed": 0, "brands": { "skilltrade.marketing": { "counts": { "mentioned": 0, "owned_cited": 0, "uncited_mention": 0, "third_party_candidates": 0 }, "rates": { "mentioned": null, "owned_cited": null, "uncited_mention": null, "third_party_candidates": null } } } } } }, "gemini": { "prompts_run": 0, "prompts_attempted": 3, "status": "incomplete", "cited": 0, "citation_share": null, "observed_share": null, "competitor_share": {}, "stages": { "problem": { "planned": 1, "completed": 0, "brands": { "skilltrade.marketing": { "counts": { "mentioned": 0, "owned_cited": 0, "uncited_mention": 0, "third_party_candidates": 0 }, "rates": { "mentioned": null, "owned_cited": null, "uncited_mention": null, "third_party_candidates": null } } } }, "category": { "planned": 1, "completed": 0, "brands": { "skilltrade.marketing": { "counts": { "mentioned": 0, "owned_cited": 0, "uncited_mention": 0, "third_party_candidates": 0 }, "rates": { "mentioned": null, "owned_cited": null, "uncited_mention": null, "third_party_candidates": null } } } }, "decision": { "planned": 1, "completed": 0, "brands": { "skilltrade.marketing": { "counts": { "mentioned": 0, "owned_cited": 0, "uncited_mention": 0, "third_party_candidates": 0 }, "rates": { "mentioned": null, "owned_cited": null, "uncited_mention": null, "third_party_candidates": null } } } } } } }, "detail": [ { "engine": "perplexity", "rows": [ { "question_id": "measure-problem", "pair_id": "measure-problem", "repeat": 1, "arm": null, "stage": "problem", "prompt": "How can a B2B marketing team tell whether AI answers name its brand?", "status": "error", "requested_model": "perplexity/sonar-pro", "provider": "openrouter", "observed_at": "2026-09-10T15:11:22.677266+00:00", "error": "HTTP 404" }, { "question_id": "measure-category", "pair_id": "measure-category", "repeat": 1, "arm": null, "stage": "category", "prompt": "What kinds of tools can a B2B marketing team use to track citations in AI answers?", "status": "error", "requested_model": "perplexity/sonar-pro", "provider": "openrouter", "observed_at": "2026-09-10T15:11:22.807233+00:00", "error": "HTTP 404" }, { "question_id": "measure-decision", "pair_id": "measure-decision", "repeat": 1, "arm": null, "stage": "decision", "prompt": "Which free, open-source tools can a B2B marketing team use to track whether AI answers cite its website?", "status": "error", "requested_model": "perplexity/sonar-pro", "provider": "openrouter", "observed_at": "2026-09-10T15:11:22.926069+00:00", "error": "HTTP 404" } ] }, { "engine": "openai", "rows": [ { "question_id": "measure-problem", "pair_id": "measure-problem", "repeat": 1, "arm": null, "stage": "problem", "prompt": "How can a B2B marketing team tell whether AI answers name its brand?", "status": "ok", "requested_model": "openai/gpt-5.2:online", "provider": "openrouter", "observed_at": "2026-09-10T15:11:23.090784+00:00", "raw_response": { "model": "openai/gpt-5.2", "choices": [ { "finish_reason": "stop", "message": { "role": "assistant", "content": "You can treat \u201cDo AI answers name us?\u201d as a measurement problem: define the questions where you *should* appear, sample answers across major AI surfaces, and track **brand mention rate + position + context** over time.\n\n## 1) Define *where you deserve to be named*\nBuild a \u201cprompt set\u201d (like an SEO keyword set) that mirrors real B2B discovery:\n\n- **Category discovery:** \u201cbest *[category]* platforms for *[use case]*\u201d\n- **Problem-first:** \u201chow to reduce *[pain]* in *[function]*\u201d\n- **Comparison / shortlist:** \u201calternatives to *[competitor]*\u201d, \u201c*[your brand]* vs *[competitor]*\u201d\n- **Integration/stack:** \u201ctools that integrate with *[system]* for *[job]*\u201d\n- **Industry-specific:** \u201c*[category]* for healthcare / fintech / manufacturing\u201d\n- **Budget/segment:** \u201centerprise *[category]*\u201d, \u201cmid-market *[category]*\u201d\n- **Geo/compliance:** \u201cSOC 2 compliant *[category]* vendors\u201d\n\nMake each prompt unambiguous and repeatable (same wording, same constraints).\n\n## 2) Test across the AI surfaces that matter (and separate \u201cmodes\u201d)\nBrand appearance varies a lot by product and mode, so test at least:\n\n- **Chat-style models** (general answers)\n- **Search-grounded answers** (web-cited / browsing / \u201csearch mode\u201d)\n- **Enterprise copilots** (often grounded in Microsoft/Google ecosystems)\n- **Citation engines** (where sources are shown)\n\nWithin a single surface, also test variants like:\n- \u201cGive me 5 vendors\u201d vs \u201cgive me 10\u201d\n- \u201cInclude pros/cons\u201d (often changes which brands appear)\n- \u201cOnly include vendors with *X* feature\u201d (filters can drop you)\n\n## 3) Track 5 core metrics (the \u201cAI Share of Voice\u201d stack)\nFor every prompt \u00d7 surface \u00d7 run, capture:\n\n1. **Mention Rate (MR):** % of answers that include your brand at least once \n2. **Position / Rank:** when a list is provided, where you appear (1st/2nd/3rd\u2026) \n3. **Context Quality:** are you recommended, neutral, or mentioned as an example/non-recommended?\n4. **Accuracy:** are key claims about you correct (features, pricing model, target segment)?\n5. **Competitor Co-mentions:** which brands appear when you do/don\u2019t\n\nRoll these up into an internal KPI like:\n- **AI SOV = (weighted mention rate) + (weighted top-3 rate) \u2013 (weighted inaccuracies)**\n\n## 4) Make it statistically stable (AI answers drift)\nOne-off checks are misleading. To reduce randomness:\n\n- **Run each prompt multiple times** (e.g., 5\u201320 runs per surface) at different times/days\n- Use **fixed settings** when possible (same model, temperature, region)\n- Record metadata: date/time, model name/version (when shown), mode (search vs non-search)\n\nThen trend weekly/monthly.\n\n## 5) Automate collection so it\u2019s not a manual project\nPractical approaches, from simplest to most robust:\n\n- **Spreadsheet + light manual sampling** (good for a pilot)\n- **Prompt harness** (internal script) that:\n - submits prompts via available APIs or approved automation,\n - stores raw outputs,\n - extracts brand mentions via rules + NER,\n - scores rank/sentiment with a second pass.\n- **Third-party \u201cAI visibility / GEO\u201d tools** (if you have budget), but still validate with your own prompt set.\n\n## 6) Diagnose *why* you\u2019re not being named\nWhen your mention rate is low, classify the failure:\n\n- **Knowledge gap:** model doesn\u2019t \u201cknow\u201d you in that category\n- **Category association issue:** you\u2019re known, but not tied to the category/use case prompt\n- **Authority gap:** competitors are more frequently cited/linked/mentioned on the web sources the AI leans on\n- **Ambiguity:** brand name overlaps with common words/other entities\n- **Eligibility mismatch:** prompt constraints exclude you (e.g., \u201copen-source\u201d, \u201cHIPAA\u201d, \u201cunder $X\u201d)\n\nThis classification matters because the fix differs.\n\n## 7) Add two \u201creality checks\u201d marketers often miss\n- **Separate \u201cnamed\u201d from \u201cclickable.\u201d** Some AI answers name brands without sending traffic. Your KPI should be \u201cnamed in recommendations,\u201d not just referral visits.\n- **Test with your real ICP language.** If customers describe the problem differently than you do, the AI will mirror the customer phrasing and may not connect it to your positioning.\n\n---\n\n### Quick starting template (you can use tomorrow)\n1. Pick **30 prompts** (10 category, 10 problem-first, 10 competitor/comparison). \n2. Test on **3 surfaces** your buyers use most. \n3. Run **5 repeats** each \u2192 30 \u00d7 3 \u00d7 5 = 450 samples. \n4. Score: mention (Y/N), rank, sentiment/context, inaccuracies. \n5. Baseline your AI SOV and re-run monthly.\n\nIf you tell me your category, top 5 competitors, and your ICP (enterprise/mid-market + vertical), I can propose a high-signal 30\u201350 prompt set and a simple scoring rubric tailored to your space.", "refusal": null, "annotations": [] } } ] }, "answer": "You can treat \u201cDo AI answers name us?\u201d as a measurement problem: define the questions where you *should* appear, sample answers across major AI surfaces, and track **brand mention rate + position + context** over time.\n\n## 1) Define *where you deserve to be named*\nBuild a \u201cprompt set\u201d (like an SEO keyword set) that mirrors real B2B discovery:\n\n- **Category discovery:** \u201cbest *[category]* platforms for *[use case]*\u201d\n- **Problem-first:** \u201chow to reduce *[pain]* in *[function]*\u201d\n- **Comparison / shortlist:** \u201calternatives to *[competitor]*\u201d, \u201c*[your brand]* vs *[competitor]*\u201d\n- **Integration/stack:** \u201ctools that integrate with *[system]* for *[job]*\u201d\n- **Industry-specific:** \u201c*[category]* for healthcare / fintech / manufacturing\u201d\n- **Budget/segment:** \u201centerprise *[category]*\u201d, \u201cmid-market *[category]*\u201d\n- **Geo/compliance:** \u201cSOC 2 compliant *[category]* vendors\u201d\n\nMake each prompt unambiguous and repeatable (same wording, same constraints).\n\n## 2) Test across the AI surfaces that matter (and separate \u201cmodes\u201d)\nBrand appearance varies a lot by product and mode, so test at least:\n\n- **Chat-style models** (general answers)\n- **Search-grounded answers** (web-cited / browsing / \u201csearch mode\u201d)\n- **Enterprise copilots** (often grounded in Microsoft/Google ecosystems)\n- **Citation engines** (where sources are shown)\n\nWithin a single surface, also test variants like:\n- \u201cGive me 5 vendors\u201d vs \u201cgive me 10\u201d\n- \u201cInclude pros/cons\u201d (often changes which brands appear)\n- \u201cOnly include vendors with *X* feature\u201d (filters can drop you)\n\n## 3) Track 5 core metrics (the \u201cAI Share of Voice\u201d stack)\nFor every prompt \u00d7 surface \u00d7 run, capture:\n\n1. **Mention Rate (MR):** % of answers that include your brand at least once \n2. **Position / Rank:** when a list is provided, where you appear (1st/2nd/3rd\u2026) \n3. **Context Quality:** are you recommended, neutral, or mentioned as an example/non-recommended?\n4. **Accuracy:** are key claims about you correct (features, pricing model, target segment)?\n5. **Competitor Co-mentions:** which brands appear when you do/don\u2019t\n\nRoll these up into an internal KPI like:\n- **AI SOV = (weighted mention rate) + (weighted top-3 rate) \u2013 (weighted inaccuracies)**\n\n## 4) Make it statistically stable (AI answers drift)\nOne-off checks are misleading. To reduce randomness:\n\n- **Run each prompt multiple times** (e.g., 5\u201320 runs per surface) at different times/days\n- Use **fixed settings** when possible (same model, temperature, region)\n- Record metadata: date/time, model name/version (when shown), mode (search vs non-search)\n\nThen trend weekly/monthly.\n\n## 5) Automate collection so it\u2019s not a manual project\nPractical approaches, from simplest to most robust:\n\n- **Spreadsheet + light manual sampling** (good for a pilot)\n- **Prompt harness** (internal script) that:\n - submits prompts via available APIs or approved automation,\n - stores raw outputs,\n - extracts brand mentions via rules + NER,\n - scores rank/sentiment with a second pass.\n- **Third-party \u201cAI visibility / GEO\u201d tools** (if you have budget), but still validate with your own prompt set.\n\n## 6) Diagnose *why* you\u2019re not being named\nWhen your mention rate is low, classify the failure:\n\n- **Knowledge gap:** model doesn\u2019t \u201cknow\u201d you in that category\n- **Category association issue:** you\u2019re known, but not tied to the category/use case prompt\n- **Authority gap:** competitors are more frequently cited/linked/mentioned on the web sources the AI leans on\n- **Ambiguity:** brand name overlaps with common words/other entities\n- **Eligibility mismatch:** prompt constraints exclude you (e.g., \u201copen-source\u201d, \u201cHIPAA\u201d, \u201cunder $X\u201d)\n\nThis classification matters because the fix differs.\n\n## 7) Add two \u201creality checks\u201d marketers often miss\n- **Separate \u201cnamed\u201d from \u201cclickable.\u201d** Some AI answers name brands without sending traffic. Your KPI should be \u201cnamed in recommendations,\u201d not just referral visits.\n- **Test with your real ICP language.** If customers describe the problem differently than you do, the AI will mirror the customer phrasing and may not connect it to your positioning.\n\n---\n\n### Quick starting template (you can use tomorrow)\n1. Pick **30 prompts** (10 category, 10 problem-first, 10 competitor/comparison). \n2. Test on **3 surfaces** your buyers use most. \n3. Run **5 repeats** each \u2192 30 \u00d7 3 \u00d7 5 = 450 samples. \n4. Score: mention (Y/N), rank, sentiment/context, inaccuracies. \n5. Baseline your AI SOV and re-run monthly.\n\nIf you tell me your category, top 5 competitors, and your ICP (enterprise/mid-market + vertical), I can propose a high-signal 30\u201350 prompt set and a simple scoring rubric tailored to your space.", "citations": [], "retrieved_urls": [], "brands": { "skilltrade.marketing": { "mentioned": false, "owned_cited": false, "uncited_mention": false, "mention_without_owned_citation": false, "third_party_candidates": [] } }, "cited": false, "competitors_cited": [], "urls": [] }, { "question_id": "measure-category", "pair_id": "measure-category", "repeat": 1, "arm": null, "stage": "category", "prompt": "What kinds of tools can a B2B marketing team use to track citations in AI answers?", "status": "ok", "requested_model": "openai/gpt-5.2:online", "provider": "openrouter", "observed_at": "2026-09-10T15:11:49.351476+00:00", "raw_response": { "model": "openai/gpt-5.2", "choices": [ { "finish_reason": "stop", "message": { "role": "assistant", "content": "B2B teams usually track \u201ccitations in AI answers\u201d with a stack of tools, because no single system consistently captures (1) *whether you were cited*, (2) *where the model got it*, and (3) *whether it drove traffic or pipeline*. Here are the main tool categories teams use.\n\n## 1) LLM/\u201cAI search\u201d visibility monitoring (purpose-built)\n**What they do:** Run a library of prompts, check results across assistants (and sometimes geos), record whether your brand/domain is cited, and trend it over time (share-of-voice, competitors, topic coverage).\n\n**When they\u2019re useful:** You care about *being referenced as a source* (not just mentioned), and you want repeatable measurement.\n\n**Examples (vendor landscape changes fast):**\n- \u201cLLM visibility\u201d / \u201cAI search optimization\u201d platforms and monitoring dashboards (often marketed as *GEO/LLMO* tools)\n- Some SEO platforms are adding AI Overviews / AI mode monitoring features\n\n## 2) Brand mention + media monitoring (classic PR monitoring, adapted to AI)\n**What they do:** Track brand/product/executive mentions across the web (news, blogs, forums, sometimes podcasts/transcripts). This doesn\u2019t directly prove \u201cAI citations,\u201d but it helps you:\n- Find the pages that *could* become AI-cited sources\n- See whether your PR pushes are generating the kind of coverage LLMs later quote/cite\n\n**Examples:** Brandwatch, Talkwalker, Meltwater, Mention, Google Alerts (lightweight).\n\n## 3) Backlink + \u201cwho cites my site\u201d tools (indirect but important)\n**What they do:** Track links to your content, link growth, referring domains, and which pages earn authority. Since many AI citation systems heavily favor authoritative, well-linked sources, backlink tooling is often the *leading indicator* for future AI citations.\n\n**Examples:** Ahrefs, Semrush, Moz, Majestic.\n\n## 4) Web analytics & server logs (to measure impact, not just presence)\n**What they do:** Tell you whether AI answers are actually driving visits, signups, or conversions. This helps answer: *\u201cEven if we\u2019re cited, does it matter?\u201d*\n\n**What to look for:**\n- Referrers (some assistants pass referrer info; some don\u2019t)\n- Landing pages that receive \u201cAI-like\u201d traffic spikes\n- Assisted conversions and downstream CRM attribution\n\n**Examples:** GA4, Adobe Analytics, Matomo; plus **server/CDN logs** (Cloudflare, Fastly, Akamai) for deeper referrer/user-agent patterns.\n\n## 5) Rank tracking + SERP feature monitoring (because AI citations often follow search visibility)\n**What they do:** Track rankings and SERP features that correlate with being pulled into AI answers (e.g., top results, featured snippets, \u201cbest of\u201d list pages, Q&A pages).\n\n**Examples:** Semrush / Ahrefs rank tracking, STAT, Similarweb (varies by plan/features).\n\n## 6) Custom \u201cprompt testing\u201d pipelines (DIY but very effective)\n**What they do:** Your team defines a prompt set (e.g., \u201cbest ERP for mid-market manufacturing,\u201d \u201cSOC 2 vendor checklist,\u201d etc.), runs it on a schedule, and stores:\n- The full answer\n- Any citations/links shown\n- The position/placement of your citation\n- Diffs over time (what changed)\n\n**How it\u2019s built:** Scripts + APIs (where allowed), scheduled runs, screenshot capture, and a small database/dashboard.\n**Tools used:** Python/Node, Playwright/Puppeteer (browser automation), a DB (BigQuery/Postgres), and a BI layer (Looker/Tableau).\n\n## 7) Content inventory + \u201csource readiness\u201d tooling (make your pages easier to cite)\n**What they do:** Ensure the pages you *want* cited are technically accessible and semantically clear\u2014so they\u2019re more likely to be selected as sources.\n\n**Examples:** Screaming Frog, Sitebulb; schema testing tools; content QA checklists (E-E-A-T style, author pages, dates, citations, update logs).\n\n---\n\n### Quick guidance on choosing the right mix\n- If your KPI is **\u201care we getting cited?\u201d** \u2192 start with **LLM visibility monitoring** + **DIY prompt library**.\n- If your KPI is **\u201cdid it drive revenue?\u201d** \u2192 prioritize **analytics/logs** + **CRM attribution**.\n- If your KPI is **\u201chow do we become cite-worthy?\u201d** \u2192 invest in **technical SEO + content/source readiness** + **digital PR/backlinks**.\n\nIf you tell me which assistants you care about (e.g., ChatGPT browsing, Perplexity, Gemini, \u201cAI Overviews\u201d), your industry, and 5\u201310 target queries, I can suggest a practical measurement setup (tooling + prompt library structure + KPIs) that a B2B team can run weekly.", "refusal": null, "annotations": [] } } ] }, "answer": "B2B teams usually track \u201ccitations in AI answers\u201d with a stack of tools, because no single system consistently captures (1) *whether you were cited*, (2) *where the model got it*, and (3) *whether it drove traffic or pipeline*. Here are the main tool categories teams use.\n\n## 1) LLM/\u201cAI search\u201d visibility monitoring (purpose-built)\n**What they do:** Run a library of prompts, check results across assistants (and sometimes geos), record whether your brand/domain is cited, and trend it over time (share-of-voice, competitors, topic coverage).\n\n**When they\u2019re useful:** You care about *being referenced as a source* (not just mentioned), and you want repeatable measurement.\n\n**Examples (vendor landscape changes fast):**\n- \u201cLLM visibility\u201d / \u201cAI search optimization\u201d platforms and monitoring dashboards (often marketed as *GEO/LLMO* tools)\n- Some SEO platforms are adding AI Overviews / AI mode monitoring features\n\n## 2) Brand mention + media monitoring (classic PR monitoring, adapted to AI)\n**What they do:** Track brand/product/executive mentions across the web (news, blogs, forums, sometimes podcasts/transcripts). This doesn\u2019t directly prove \u201cAI citations,\u201d but it helps you:\n- Find the pages that *could* become AI-cited sources\n- See whether your PR pushes are generating the kind of coverage LLMs later quote/cite\n\n**Examples:** Brandwatch, Talkwalker, Meltwater, Mention, Google Alerts (lightweight).\n\n## 3) Backlink + \u201cwho cites my site\u201d tools (indirect but important)\n**What they do:** Track links to your content, link growth, referring domains, and which pages earn authority. Since many AI citation systems heavily favor authoritative, well-linked sources, backlink tooling is often the *leading indicator* for future AI citations.\n\n**Examples:** Ahrefs, Semrush, Moz, Majestic.\n\n## 4) Web analytics & server logs (to measure impact, not just presence)\n**What they do:** Tell you whether AI answers are actually driving visits, signups, or conversions. This helps answer: *\u201cEven if we\u2019re cited, does it matter?\u201d*\n\n**What to look for:**\n- Referrers (some assistants pass referrer info; some don\u2019t)\n- Landing pages that receive \u201cAI-like\u201d traffic spikes\n- Assisted conversions and downstream CRM attribution\n\n**Examples:** GA4, Adobe Analytics, Matomo; plus **server/CDN logs** (Cloudflare, Fastly, Akamai) for deeper referrer/user-agent patterns.\n\n## 5) Rank tracking + SERP feature monitoring (because AI citations often follow search visibility)\n**What they do:** Track rankings and SERP features that correlate with being pulled into AI answers (e.g., top results, featured snippets, \u201cbest of\u201d list pages, Q&A pages).\n\n**Examples:** Semrush / Ahrefs rank tracking, STAT, Similarweb (varies by plan/features).\n\n## 6) Custom \u201cprompt testing\u201d pipelines (DIY but very effective)\n**What they do:** Your team defines a prompt set (e.g., \u201cbest ERP for mid-market manufacturing,\u201d \u201cSOC 2 vendor checklist,\u201d etc.), runs it on a schedule, and stores:\n- The full answer\n- Any citations/links shown\n- The position/placement of your citation\n- Diffs over time (what changed)\n\n**How it\u2019s built:** Scripts + APIs (where allowed), scheduled runs, screenshot capture, and a small database/dashboard.\n**Tools used:** Python/Node, Playwright/Puppeteer (browser automation), a DB (BigQuery/Postgres), and a BI layer (Looker/Tableau).\n\n## 7) Content inventory + \u201csource readiness\u201d tooling (make your pages easier to cite)\n**What they do:** Ensure the pages you *want* cited are technically accessible and semantically clear\u2014so they\u2019re more likely to be selected as sources.\n\n**Examples:** Screaming Frog, Sitebulb; schema testing tools; content QA checklists (E-E-A-T style, author pages, dates, citations, update logs).\n\n---\n\n### Quick guidance on choosing the right mix\n- If your KPI is **\u201care we getting cited?\u201d** \u2192 start with **LLM visibility monitoring** + **DIY prompt library**.\n- If your KPI is **\u201cdid it drive revenue?\u201d** \u2192 prioritize **analytics/logs** + **CRM attribution**.\n- If your KPI is **\u201chow do we become cite-worthy?\u201d** \u2192 invest in **technical SEO + content/source readiness** + **digital PR/backlinks**.\n\nIf you tell me which assistants you care about (e.g., ChatGPT browsing, Perplexity, Gemini, \u201cAI Overviews\u201d), your industry, and 5\u201310 target queries, I can suggest a practical measurement setup (tooling + prompt library structure + KPIs) that a B2B team can run weekly.", "citations": [], "retrieved_urls": [], "brands": { "skilltrade.marketing": { "mentioned": false, "owned_cited": false, "uncited_mention": false, "mention_without_owned_citation": false, "third_party_candidates": [] } }, "cited": false, "competitors_cited": [], "urls": [] }, { "question_id": "measure-decision", "pair_id": "measure-decision", "repeat": 1, "arm": null, "stage": "decision", "prompt": "Which free, open-source tools can a B2B marketing team use to track whether AI answers cite its website?", "status": "ok", "requested_model": "openai/gpt-5.2:online", "provider": "openrouter", "observed_at": "2026-09-10T15:12:20.189939+00:00", "raw_response": { "model": "openai/gpt-5.2", "choices": [ { "finish_reason": "stop", "message": { "role": "assistant", "content": "Below are **free, open-source** options that specifically help you **measure whether LLM/AI \u201canswer engines\u201d cite (link to) your website**\u2014typically by running a fixed set of prompts on a schedule, extracting the **cited URLs/domains**, and trending results over time.\n\n## Purpose-built open-source \u201cAI citation / visibility trackers\u201d\n\n- **Aperture (MIT)** \u2014 Self-hosted \u201cAI visibility infrastructure\u201d to track whether AI models recommend/cite your brand for chosen queries, plus competitor comparisons and time-series tracking. ([github.com](https://github.com/anyin-ai/aperture)) \n *Good when you want a team-friendly, self-hosted app rather than a CLI script.*\n\n- **geo-tracker (MIT)** \u2014 Lightweight CLI that runs prompts across multiple engines (via OpenRouter \u201conline\u201d models) and logs citations to a local **SQLite ledger** so you can track citation rate week-over-week. ([github.com](https://github.com/onblueroses/geo-tracker)) \n *Good when you want something simple, auditable, and scriptable (cron/GitHub Actions).*\n\n- **Ansvisor (MIT)** \u2014 Open-source \u201cAI Search Intelligence Platform\u201d positioned for tracking visibility/citations across multiple answer engines. ([github.com](https://github.com/ansvisor/ansvisor)) \n *Good if you want a broader \u201cplatform\u201d approach (visibility + analysis workflows).*\n\n- **AEO Tracker (MIT)** \u2014 \u201cRead-only\u201d citation tracker that queries multiple engines (Claude/Perplexity/OpenAI/Gemini) and records whether a target domain appears in citations, storing runs in **SQLite** for trend reports. ([github.com](https://github.com/major-matters/aeo-tracker)) \n *Good when you want a minimal, engineer-friendly measurement instrument.*\n\n- **Cite.me.in (AGPL-3.0, archived)** \u2014 Tracks when/where AI platforms cite your site, storing citations per query over time; repo is archived but still usable/forkable. ([github.com](https://github.com/cite-me-in/cite.me.in)) \n *Good as reference code or if you\u2019re comfortable adopting an archived project.*\n\n## Google AI Overviews / AI Mode (browser-capture approach)\nFor Google\u2019s AI surfaces, APIs are limited, so **browser instrumentation** is common:\n\n- **Brand AI Rank Tracker (Chrome extension)** \u2014 Captures **Google AI Overviews / AI Mode** payloads (including citations/grounding hosts) from searches you run, and logs \u201cvisibility lanes\u201d (visible citation vs grounded-only, etc.). ([github.com](https://github.com/vijaychauhanseo/brand-ai-rank-tracker)) \n *Good when your main goal is \u201cDoes Google\u2019s AI experience cite us?\u201d*\n\n## Small, hackable scripts (good starting point)\n- **SkillTrade `citation_meter.py` (open-source script)** \u2014 Runs a fixed question set across assistants, extracts cited URLs/domains, and trends results; designed to be resilient to minor API response-shape changes. ([skilltrade.marketing](https://www.skilltrade.marketing/tools/)) \n *Good if you want something you can read/modify in an afternoon.*\n\n---\n\n## Practical notes (so expectations are right)\n- \u201cFree & open-source\u201d usually means **the software is free**, but **the underlying model/search APIs cost money** (often \u201ccents per run\u201d) unless you point it at a local model that can do web retrieval. (Several of the tools above are explicitly BYOK\u2014bring your own keys.) ([github.com](https://github.com/anyin-ai/aperture))\n- Results are **non-deterministic** (the same query may cite different sources on different days), so the winning move is running a **consistent prompt set** on a schedule and tracking trends. ([github.com](https://github.com/major-matters/aeo-tracker))\n\nIf you tell me (1) which engines you care about most (ChatGPT vs Perplexity vs Google AIO/AIM), (2) whether you prefer **self-hosted UI** vs **CLI**, and (3) how many prompts you want to track (10? 100? 1000?), I can recommend the best 1\u20132 picks and a simple rollout plan.", "refusal": null, "annotations": [ { "type": "url_citation", "url_citation": { "url": "https://github.com/anyin-ai/aperture", "title": "GitHub - anyin-ai/aperture: Open-source AI visibility monitoring and analytics. Track how your brand appears across ChatGPT, Perplexity, Google AI Overviews, and other LLMs. BYOK. Self-hosted. Free alternative to Profound and Peec AI. \u00b7 GitHub", "start_index": 529, "end_index": 581 } }, { "type": "url_citation", "url_citation": { "url": "https://github.com/onblueroses/geo-tracker", "title": "GitHub - onblueroses/geo-tracker: Measure how often LLM answer engines (Perplexity, ChatGPT, Claude, Gemini) cite your URLs. SQLite ledger, OpenRouter-backed, append-only. \u00b7 GitHub", "start_index": 879, "end_index": 937 } }, { "type": "url_citation", "url_citation": { "url": "https://github.com/ansvisor/ansvisor", "title": "GitHub - ansvisor/ansvisor: Open-source AI Search Intelligence Platform \u2014 track, analyze, and improve AI visibility, citations, prompts, competitors, and content opportunities across ChatGPT, Claude, Gemini, Google AI Overviews, Google AI Mode, Perplexity, Grok, and Microsoft Copilot. \u00b7 GitHub", "start_index": 1177, "end_index": 1229 } }, { "type": "url_citation", "url_citation": { "url": "https://github.com/major-matters/aeo-tracker", "title": "GitHub - major-matters/aeo-tracker: Which answer engines cite your URLs. Read-only citation tracker across Claude, Perplexity, OpenAI, and Gemini. Bring your own keys. \u00b7 GitHub", "start_index": 1542, "end_index": 1602 } }, { "type": "url_citation", "url_citation": { "url": "https://github.com/cite-me-in/cite.me.in", "title": "GitHub - cite-me-in/cite.me.in: Monitor your brand's visibility in AI-generated responses. \u00b7 GitHub", "start_index": 1852, "end_index": 1908 } }, { "type": "url_citation", "url_citation": { "url": "https://github.com/vijaychauhanseo/brand-ai-rank-tracker", "title": "GitHub - vijaychauhanseo/brand-ai-rank-tracker: Open-source Chrome extension that tracks your brand in Google AI Overviews and AI Mode. Lane model, DEJAN lens, response forensics, MCP-ready. 100% local. \u00b7 GitHub", "start_index": 2374, "end_index": 2446 } }, { "type": "url_citation", "url_citation": { "url": "https://www.skilltrade.marketing/tools/", "title": "AI Citation Tracking: Free Tools to Measure AI Visibility", "start_index": 2786, "end_index": 2851 } } ] } } ] }, "answer": "Below are **free, open-source** options that specifically help you **measure whether LLM/AI \u201canswer engines\u201d cite (link to) your website**\u2014typically by running a fixed set of prompts on a schedule, extracting the **cited URLs/domains**, and trending results over time.\n\n## Purpose-built open-source \u201cAI citation / visibility trackers\u201d\n\n- **Aperture (MIT)** \u2014 Self-hosted \u201cAI visibility infrastructure\u201d to track whether AI models recommend/cite your brand for chosen queries, plus competitor comparisons and time-series tracking. ([github.com](https://github.com/anyin-ai/aperture)) \n *Good when you want a team-friendly, self-hosted app rather than a CLI script.*\n\n- **geo-tracker (MIT)** \u2014 Lightweight CLI that runs prompts across multiple engines (via OpenRouter \u201conline\u201d models) and logs citations to a local **SQLite ledger** so you can track citation rate week-over-week. ([github.com](https://github.com/onblueroses/geo-tracker)) \n *Good when you want something simple, auditable, and scriptable (cron/GitHub Actions).*\n\n- **Ansvisor (MIT)** \u2014 Open-source \u201cAI Search Intelligence Platform\u201d positioned for tracking visibility/citations across multiple answer engines. ([github.com](https://github.com/ansvisor/ansvisor)) \n *Good if you want a broader \u201cplatform\u201d approach (visibility + analysis workflows).*\n\n- **AEO Tracker (MIT)** \u2014 \u201cRead-only\u201d citation tracker that queries multiple engines (Claude/Perplexity/OpenAI/Gemini) and records whether a target domain appears in citations, storing runs in **SQLite** for trend reports. ([github.com](https://github.com/major-matters/aeo-tracker)) \n *Good when you want a minimal, engineer-friendly measurement instrument.*\n\n- **Cite.me.in (AGPL-3.0, archived)** \u2014 Tracks when/where AI platforms cite your site, storing citations per query over time; repo is archived but still usable/forkable. ([github.com](https://github.com/cite-me-in/cite.me.in)) \n *Good as reference code or if you\u2019re comfortable adopting an archived project.*\n\n## Google AI Overviews / AI Mode (browser-capture approach)\nFor Google\u2019s AI surfaces, APIs are limited, so **browser instrumentation** is common:\n\n- **Brand AI Rank Tracker (Chrome extension)** \u2014 Captures **Google AI Overviews / AI Mode** payloads (including citations/grounding hosts) from searches you run, and logs \u201cvisibility lanes\u201d (visible citation vs grounded-only, etc.). ([github.com](https://github.com/vijaychauhanseo/brand-ai-rank-tracker)) \n *Good when your main goal is \u201cDoes Google\u2019s AI experience cite us?\u201d*\n\n## Small, hackable scripts (good starting point)\n- **SkillTrade `citation_meter.py` (open-source script)** \u2014 Runs a fixed question set across assistants, extracts cited URLs/domains, and trends results; designed to be resilient to minor API response-shape changes. ([skilltrade.marketing](https://www.skilltrade.marketing/tools/)) \n *Good if you want something you can read/modify in an afternoon.*\n\n---\n\n## Practical notes (so expectations are right)\n- \u201cFree & open-source\u201d usually means **the software is free**, but **the underlying model/search APIs cost money** (often \u201ccents per run\u201d) unless you point it at a local model that can do web retrieval. (Several of the tools above are explicitly BYOK\u2014bring your own keys.) ([github.com](https://github.com/anyin-ai/aperture))\n- Results are **non-deterministic** (the same query may cite different sources on different days), so the winning move is running a **consistent prompt set** on a schedule and tracking trends. ([github.com](https://github.com/major-matters/aeo-tracker))\n\nIf you tell me (1) which engines you care about most (ChatGPT vs Perplexity vs Google AIO/AIM), (2) whether you prefer **self-hosted UI** vs **CLI**, and (3) how many prompts you want to track (10? 100? 1000?), I can recommend the best 1\u20132 picks and a simple rollout plan.", "citations": [ { "url": "https://github.com/anyin-ai/aperture", "passage": "([github.com](https://github.com/anyin-ai/aperture))", "title": "GitHub - anyin-ai/aperture: Open-source AI visibility monitoring and analytics. Track how your brand appears across ChatGPT, Perplexity, Google AI Overviews, and other LLMs. BYOK. Self-hosted. Free alternative to Profound and Peec AI. \u00b7 GitHub" }, { "url": "https://github.com/onblueroses/geo-tracker", "passage": "([github.com](https://github.com/onblueroses/geo-tracker))", "title": "GitHub - onblueroses/geo-tracker: Measure how often LLM answer engines (Perplexity, ChatGPT, Claude, Gemini) cite your URLs. SQLite ledger, OpenRouter-backed, append-only. \u00b7 GitHub" }, { "url": "https://github.com/ansvisor/ansvisor", "passage": "([github.com](https://github.com/ansvisor/ansvisor))", "title": "GitHub - ansvisor/ansvisor: Open-source AI Search Intelligence Platform \u2014 track, analyze, and improve AI visibility, citations, prompts, competitors, and content opportunities across ChatGPT, Claude, Gemini, Google AI Overviews, Google AI Mode, Perplexity, Grok, and Microsoft Copilot. \u00b7 GitHub" }, { "url": "https://github.com/major-matters/aeo-tracker", "passage": "([github.com](https://github.com/major-matters/aeo-tracker))", "title": "GitHub - major-matters/aeo-tracker: Which answer engines cite your URLs. Read-only citation tracker across Claude, Perplexity, OpenAI, and Gemini. Bring your own keys. \u00b7 GitHub" }, { "url": "https://github.com/cite-me-in/cite.me.in", "passage": "([github.com](https://github.com/cite-me-in/cite.me.in))", "title": "GitHub - cite-me-in/cite.me.in: Monitor your brand's visibility in AI-generated responses. \u00b7 GitHub" }, { "url": "https://github.com/vijaychauhanseo/brand-ai-rank-tracker", "passage": "([github.com](https://github.com/vijaychauhanseo/brand-ai-rank-tracker))", "title": "GitHub - vijaychauhanseo/brand-ai-rank-tracker: Open-source Chrome extension that tracks your brand in Google AI Overviews and AI Mode. Lane model, DEJAN lens, response forensics, MCP-ready. 100% local. \u00b7 GitHub" }, { "url": "https://www.skilltrade.marketing/tools/", "passage": "([skilltrade.marketing](https://www.skilltrade.marketing/tools/))", "title": "AI Citation Tracking: Free Tools to Measure AI Visibility" } ], "retrieved_urls": [], "brands": { "skilltrade.marketing": { "mentioned": true, "owned_cited": true, "uncited_mention": false, "mention_without_owned_citation": false, "third_party_candidates": [] } }, "cited": true, "competitors_cited": [], "urls": [ "https://github.com/ansvisor/ansvisor", "https://github.com/anyin-ai/aperture", "https://github.com/cite-me-in/cite.me.in", "https://github.com/major-matters/aeo-tracker", "https://github.com/onblueroses/geo-tracker", "https://github.com/vijaychauhanseo/brand-ai-rank-tracker", "https://www.skilltrade.marketing/tools/" ] } ] }, { "engine": "anthropic", "rows": [ { "question_id": "measure-problem", "pair_id": "measure-problem", "repeat": 1, "arm": null, "stage": "problem", "prompt": "How can a B2B marketing team tell whether AI answers name its brand?", "status": "error", "requested_model": "anthropic/claude-sonnet-5:online", "provider": "openrouter", "observed_at": "2026-09-10T15:13:08.573672+00:00", "raw_response": { "model": "anthropic/claude-sonnet-5", "choices": [ { "finish_reason": "stop", "message": { "role": "assistant", "content": "B2B marketing teams can determine whether AI answers (ChatGPT, Perplexity, Gemini, Claude, Copilot, Google AI Overviews, etc.) name their brand through a combination of manual auditing and automated monitoring. Here's how the process typically works:\n\n## Why This Matters\nThe core problem is that unlike traditional search where you can at least see your rankings, AI chatbots operate as black boxes\u2014you have no idea if they're recommending your product, mentioning your competitors exclusively, or ignoring your category entirely. This creates real business risk: Google AI Mode has reportedly reached 75 million daily active users, and approximately 93% of those sessions end without a single click to any website. If your brand isn't named in the answer itself, you may have zero visibility into that customer interaction.\n\n## 1. Start With a Manual Baseline Audit\nBefore investing in tools, teams should run initial manual queries on each platform using prompts your ideal customer would actually ask, such as \"What are the best [your category] tools?\" or \"How do I solve [core problem your product addresses]?\" Then check whether your brand appears, where it appears in the response, and how it's described\u2014this manual baseline audit takes a few hours but is genuinely eye-opening.\n\nTeams should decide which platforms matter most for their buyers. ChatGPT is the most widely used AI assistant globally, with both a knowledge cutoff model and a web-browsing version, particularly popular among professionals and B2B researchers. Meanwhile Claude is known for nuanced, detailed responses and is increasingly adopted by enterprise teams and developers, though it often pulls from its training data rather than real-time web retrieval. Perplexity is a search-focused AI that actively retrieves and cites live web sources. Notably, B2B audiences researching software tools tend to skew heavily toward ChatGPT and Perplexity.\n\n## 2. Build a Structured Prompt Library\nNot all prompts surface brand mentions equally. There are generally three main prompt categories that drive brand mentions in commercial contexts. The first is the \"best tools\" format: \"What are the best SEO tools for a B2B SaaS company?\" or \"What's the best CRM for enterprise sales teams?\" which tend to surface category leaders and well-documented tools. The second is the comparison format: \"Compare [Tool A] versus alternatives for [use case]\" or \"What are the main differences between [Tool A] and similar products?\"\u2014these prompts reveal how AI models position your brand relative to competitors.\n\nVolume matters for statistical reliability: research shows you need volume (60-100 runs per prompt) to produce statistically meaningful visibility data, so daily automated tracking that accumulates over time gives you the most reliable picture.\n\n## 3. Move to Automated, Multi-Engine Monitoring\nManual checks don't scale. As one guide notes, manual tracking is impossible at scale\u2014use platforms that automate querying, parsing, and reporting. Key considerations for automated tracking:\n\n- **Frequency**: Weekly or biweekly checks capture changes promptly.\n- **Multi-platform coverage**: Multi-platform coverage is non-negotiable in 2026. Different AI engines pull from different sources and have different training data cutoffs. A brand might appear in Perplexity consistently but be completely absent from ChatGPT free-tier responses. Tracking a single engine gives an incomplete picture.\n- **Position matters, not just presence**: Models like Perplexity, ChatGPT, and Grok mention brands near the start of responses (median rank 1), while Claude tends to mention brands later (median rank 3).\n\nDedicated AI visibility platforms (e.g., Otterly.ai, Peec AI, Profound, Evertune, Semrush One, and others) exist specifically for this. These tools generally work by sending hundreds of relevant prompts to ChatGPT, Perplexity, Gemini, and other engines on a recurring schedule, capturing responses, and analyzing whether your brand was mentioned, cited, or recommended.\n\n## 4. Distinguish Mentions From Citations\nIt's worth separating two related but distinct signals: AI mentions occur when your brand name appears in an AI answer, while AI citations are when AI uses your brand as a source\u2014citations are stronger authority signals.\n\n## 5. Track Share of Voice vs. Competitors\nBeyond simple presence/absence, teams should measure competitive standing: AI share of voice measures how often your brand appears vs. competitors in AI-generated answers\u2014if competitors have higher AI share of voice, they're capturing attention you're missing.\n\n## 6. Connect Mentions to Real Traffic (With Caveats)\nAnalytics tools have real limitations here. GA4 can track referral traffic from ChatGPT when users click citation links, but only about 20% of ChatGPT mentions include clickable links\u2014the other 80% of brand recommendations, descriptions, and comparisons are invisible to GA4. Perplexity is different: every citation is clickable, so Perplexity referral traffic appears cleanly in analytics.\n\n## Bottom Line\nThe practical playbook is: run a manual baseline audit across the 3-4 AI platforms your buyers actually use, build a prompt library mirroring real buyer questions, then layer in an automated tracking tool for ongoing, scalable monitoring \u2014 since as one guide puts it, a static snapshot tells you where you were; a continuous tracking cadence tells you whether your strategy is working.", "refusal": null, "annotations": [ { "type": "url_citation", "url_citation": { "url": "https://www.trysight.ai/blog/brand-tracking-in-ai-chatbots", "title": "Brand Tracking In AI Chatbots: Complete 2026 Guide - Sight AI", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.digitalapplied.com/blog/ai-visibility-tools-2026-track-brand-chatgpt-perplexity-gemini", "title": "AI Visibility Tools 2026: Track Your Brand Across LLMs", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.trysight.ai/blog/track-ai-chatbot-brand-mentions", "title": "How to Track AI Chatbot Brand Mentions Effectively", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.trysight.ai/blog/track-ai-chatbot-brand-mentions", "title": "How to Track AI Chatbot Brand Mentions Effectively", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.trysight.ai/blog/track-ai-chatbot-brand-mentions", "title": "How to Track AI Chatbot Brand Mentions Effectively", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.trysight.ai/blog/track-ai-chatbot-brand-mentions", "title": "How to Track AI Chatbot Brand Mentions Effectively", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.trysight.ai/blog/track-ai-chatbot-brand-mentions", "title": "How to Track AI Chatbot Brand Mentions Effectively", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.trysight.ai/blog/track-ai-chatbot-brand-mentions", "title": "How to Track AI Chatbot Brand Mentions Effectively", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.trysight.ai/blog/brand-mention-tracking-across-chatbots", "title": "Brand Mention Tracking Across Chatbots: A Full Guide", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.trysight.ai/blog/brand-mention-tracking-across-chatbots", "title": "Brand Mention Tracking Across Chatbots: A Full Guide", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.trysight.ai/blog/brand-mention-tracking-across-chatbots", "title": "Brand Mention Tracking Across Chatbots: A Full Guide", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.getpassionfruit.com/blog/how-to-monitor-your-brand-across-chatgpt-perplexity-and-ai-search", "title": "How to Monitor Your Brand in ChatGPT, Perplexity & AI Search (2026 Guide)", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://rocketblue.ai/articles/tracking-brand-mentions-in-ai-chatbots-a-comprehensive-guide-to-monitoring-brand-presence-in-chatgpt-responses-feb-2026-data/", "title": "Tracking Brand Mentions in AI Chatbots: A Comprehensive Guide to Monitoring Brand Presence in ChatGPT Responses (Feb 2026 data) - rocketblue Articles", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://rocketblue.ai/articles/tracking-brand-mentions-in-ai-chatbots-a-comprehensive-guide-to-monitoring-brand-presence-in-chatgpt-responses-feb-2026-data/", "title": "Tracking Brand Mentions in AI Chatbots: A Comprehensive Guide to Monitoring Brand Presence in ChatGPT Responses (Feb 2026 data) - rocketblue Articles", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://pallasai.io/blog/track-brand-mentions-in-chatgpt-perplexity-tools-guide", "title": "Track Brand Mentions in ChatGPT & Perplexity: Tools Guide | PallasAI Blog", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://rocketblue.ai/articles/tracking-brand-mentions-in-ai-chatbots-a-comprehensive-guide-to-monitoring-brand-presence-in-chatgpt-responses-feb-2026-data/", "title": "Tracking Brand Mentions in AI Chatbots: A Comprehensive Guide to Monitoring Brand Presence in ChatGPT Responses (Feb 2026 data) - rocketblue Articles", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://pallasai.io/blog/track-brand-mentions-in-chatgpt-perplexity-tools-guide", "title": "Track Brand Mentions in ChatGPT & Perplexity: Tools Guide | PallasAI Blog", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.frizerly.com/track-brand-mentions-ai-answers.html", "title": "Best Tool to Track & Boost Your Brand's Visibility in AI Search 2026 | ChatGPT, Perplexity, Gemini, Claude, Grok | Frizerly", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.frizerly.com/track-brand-mentions-ai-answers.html", "title": "Best Tool to Track & Boost Your Brand's Visibility in AI Search 2026 | ChatGPT, Perplexity, Gemini, Claude, Grok | Frizerly", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.getpassionfruit.com/blog/how-to-monitor-your-brand-across-chatgpt-perplexity-and-ai-search", "title": "How to Monitor Your Brand in ChatGPT, Perplexity & AI Search (2026 Guide)", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.getpassionfruit.com/blog/how-to-monitor-your-brand-across-chatgpt-perplexity-and-ai-search", "title": "How to Monitor Your Brand in ChatGPT, Perplexity & AI Search (2026 Guide)", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.trysight.ai/blog/brand-mention-tracking-across-chatbots", "title": "Brand Mention Tracking Across Chatbots: A Full Guide", "start_index": 0, "end_index": 0 } } ] } } ] }, "error": "EvidenceError" }, { "question_id": "measure-category", "pair_id": "measure-category", "repeat": 1, "arm": null, "stage": "category", "prompt": "What kinds of tools can a B2B marketing team use to track citations in AI answers?", "status": "error", "requested_model": "anthropic/claude-sonnet-5:online", "provider": "openrouter", "observed_at": "2026-09-10T15:13:35.745726+00:00", "raw_response": { "model": "anthropic/claude-sonnet-5", "choices": [ { "finish_reason": "stop", "message": { "role": "assistant", "content": "# Tools for Tracking Citations in AI Answers\n\nThis category is often called \"AI visibility tracking,\" \"AI citation tracking,\" or \"Generative Engine Optimization (GEO)\" tools. Here's what's available for B2B marketing teams:\n\n## What These Tools Actually Do\n\nAt their core, AI visibility tracking tools monitor how brands appear in answers generated by ChatGPT, Perplexity, Claude, Gemini, Google AI Overviews, and other AI engines. More specifically, AI visibility platforms monitor citation frequency, source attribution, sentiment, and competitor share of voice across multiple LLMs.\n\nAn important distinction many of these platforms make is between a mention and a true citation: the first distinction is brand mention vs citation \u2014 a brand mention means the answer names your company, product, or domain, while a citation means the answer includes a linked source, source panel, or cited URL connected to your website.\n\n## Key Tool Categories & Examples\n\n**Enterprise / Comprehensive Platforms**\n- Profound is an AI answer-engine analytics platform that tracks brand mentions, share of voice, and competitive positioning across multiple LLM models, designed for enterprise marketing teams who need daily competitive benchmarking and trend alerts\n- Other enterprise-focused names that repeatedly surface include AthenaHQ, Semrush's AI toolkit, and Evertune, which fill the gap left by traditional rank trackers by monitoring how ChatGPT, Perplexity, Gemini, and Claude reference your brand\n\n**Mid-Market / Accessible Options**\n- OtterlyAI, at $29/month, is positioned as the most accessible serious option, serving 20,000+ users, with multi-platform coverage that's attractive for agencies managing multiple brands\n- Xfunnel AI is purpose-built for tracking brand and URL citations inside LLM responses, running automated query sets across multiple AI platforms and logging which URLs get cited, at what frequency, and starting at $79/month with a free tier\n\n**Tools with Distinctive Features**\n- Peec AI stands out for separating \"used\" from \"cited\" sources: its defining feature is the Used vs. Cited distinction \u2014 most tools collapse these into a single \"mention\" metric, which obscures the real problem, but Peec separates content that informed an AI answer from content explicitly linked as a source\n- Xfunnel also tracks citation sentiment, flagging cases where AI mentions your brand in a negative or ambiguous context\n\n**Platforms Combining Tracking + Content Action**\nSome newer tools pair diagnosis with remediation. For example, DeepSmith is an AI search analytics and content production platform in one \u2014 it tracks how AI engines answer questions about your brand, finds the gaps where you are invisible or losing, and produces the on-brand articles to close those gaps, all from the same data. Similarly, Frase tracks your brand across ChatGPT, Perplexity, Claude, Gemini, and Google AI, then feeds that signal straight into research, drafting, scoring, and publishing \u2014 turning a lost citation into a brief, a draft in your brand voice, and a published fix.\n\n## Why This Matters for B2B Specifically\n\nThe stakes are high for B2B buyer research: 72% of B2B buyers now encounter AI Overviews during their research process, and 90% of those buyers click cited sources to verify information. Additionally, ranking well in traditional search no longer guarantees AI citation \u2014 an Ahrefs citation analysis of 863,000 keywords found that only 38% of AI Overview citations now come from top-10 ranking positions, down from 76% in July 2024, meaning ranking alone no longer predicts citation.\n\nB2B-specific use cases these tools address include: comparison prompts (like Salesforce vs HubSpot), integration queries, and security/compliance questions that traditional SEO tools miss \u2014 with platforms like Omnia, Profound, and AthenaHQ tracking competitor displacement patterns to show when prospects research a category but competitors get cited instead.\n\n## What to Look for When Choosing\n\nBased on common evaluation criteria across these tools, prioritize:\n1. **Engine coverage** \u2014 whether it tracks the AI engines your buyers actually use: ChatGPT, Perplexity, Gemini, Claude, and Google AI Mode at minimum\n2. **Mention vs. citation clarity** \u2014 tools that distinguish being *named* from being *linked as a source*\n3. **Competitive benchmarking** \u2014 seeing which competitors get cited instead of you\n4. **Actionability** \u2014 whether the platform just reports data or helps you close content gaps\n5. **Sentiment/context tracking** \u2014 understanding *how* you're being described, not just *whether*\n\nA quick caveat worth noting: no AI visibility checker can fully track every possible AI answer, since AI outputs vary by prompt wording, location, account context, model version, retrieval behavior, browsing availability, and time. Treat these tools as directional trend-monitors rather than perfectly precise measurement instruments.", "refusal": null, "annotations": [ { "type": "url_citation", "url_citation": { "url": "https://www.rivo.io/blog/ai-visibility-tracking-tools", "title": "Rivo | Best AI Visibility Tracking Tools (2026)", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.visiblie.com/blog/best-ai-visibility-tools", "title": "8 Best AI Visibility Tools Tested & Ranked [2026] | Visiblie", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://seeklab.io/blog/best-ai-visibility-checkers-in-2026-which-tools-actually-track-chatgpt-and-perplexity-citations/", "title": "Best AI Visibility Checkers In 2026: Which Tools Actually Track ChatGPT And Perplexity Citations", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.visibilitystack.ai/signals/listicle/ai-brand-monitoring-tools", "title": "Top AI Brand Monitoring Tools 2026 | Citation Tracking", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.digitalapplied.com/blog/ai-visibility-tools-2026-track-brand-chatgpt-perplexity-gemini", "title": "AI Visibility Tools 2026: Track Your Brand Across LLMs", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://meev.ai/articles/top-ai-citation-tracking-tools-for-b2b-marketing", "title": "Top AI Citation Tracking Tools for B2B Marketing 2026 | Meev", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.stackmatix.com/blog/ai-citation-tracking-tools", "title": "AI Citation Tracking Tools: Monitor Your Brand (2026)", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.stackmatix.com/blog/ai-citation-tracking-tools", "title": "AI Citation Tracking Tools: Monitor Your Brand (2026)", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://meev.ai/articles/top-ai-citation-tracking-tools-for-b2b-marketing", "title": "Top AI Citation Tracking Tools for B2B Marketing 2026 | Meev", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.stackmatix.com/blog/ai-citation-tracking-tools", "title": "AI Citation Tracking Tools: Monitor Your Brand (2026)", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://deepsmith.ai/blog/best-ai-citation-tracking-tools-for-b2b", "title": "AI Citation Tracking Tools for B2B Marketing Teams | DeepSmith", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.frase.io/blog/the-10-best-ai-visibility-tools-in-2026", "title": "10 Best AI Visibility Tools in 2026 (Compared + Free Checker) | Frase", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.frase.io/blog/the-10-best-ai-visibility-tools-in-2026", "title": "10 Best AI Visibility Tools in 2026 (Compared + Free Checker) | Frase", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://meev.ai/articles/top-ai-citation-tracking-tools-for-b2b-marketing", "title": "Top AI Citation Tracking Tools for B2B Marketing 2026 | Meev", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://meev.ai/articles/top-ai-citation-tracking-tools-for-b2b-marketing", "title": "Top AI Citation Tracking Tools for B2B Marketing 2026 | Meev", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.useomnia.com/blog/ai-search-monitoring-tools", "title": "AI Search Monitoring Tools 2026: The Best Platforms to Track Mentions, Citations, and Visibility", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://deepsmith.ai/blog/best-ai-citation-tracking-tools-for-b2b", "title": "AI Citation Tracking Tools for B2B Marketing Teams | DeepSmith", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://seeklab.io/blog/best-ai-visibility-checkers-in-2026-which-tools-actually-track-chatgpt-and-perplexity-citations/", "title": "Best AI Visibility Checkers In 2026: Which Tools Actually Track ChatGPT And Perplexity Citations", "start_index": 0, "end_index": 0 } } ] } } ] }, "error": "EvidenceError" }, { "question_id": "measure-decision", "pair_id": "measure-decision", "repeat": 1, "arm": null, "stage": "decision", "prompt": "Which free, open-source tools can a B2B marketing team use to track whether AI answers cite its website?", "status": "error", "requested_model": "anthropic/claude-sonnet-5:online", "provider": "openrouter", "observed_at": "2026-09-10T15:14:01.726173+00:00", "raw_response": { "model": "anthropic/claude-sonnet-5", "choices": [ { "finish_reason": "stop", "message": { "role": "assistant", "content": "Here are several free, open-source tools a B2B marketing team can self-host to monitor whether AI answer engines (ChatGPT, Perplexity, Gemini, Claude, Copilot, etc.) cite their website:\n\n## Open-source, self-hosted trackers\n\n**Ansvisor** \u2014 Ansvisor is an open-source AI Visibility platform built to track website mentions, citations, competitors, and visibility across major AI answer engines. It monitors platforms including ChatGPT, Gemini, Perplexity, Claude, Microsoft Copilot, and its feature set includes real-time visibility scores across ChatGPT, Google AI Overview, Google AI Mode, Google Gemini, Perplexity, Microsoft Copilot, Grok, and Claude with weekly trend analysis, plus a citations module that lets teams discover which domains, URLs, competitors, and third-party sources AI systems cite. It's designed so you can self-host Ansvisor for free or use the managed Ansvisor Cloud.\n\n**GEO/AEO Tracker (Bright Data-based)** \u2014 This is a free, open-source AI visibility dashboard tracking 6 AI models simultaneously, using Bright Data's LLM Scrapers to query ChatGPT, Gemini, Perplexity, Grok, Copilot, and Google AI Mode, delivering structured output (citations, sources, answer text) per model, via a single API pattern. It runs tracking your brand across ChatGPT, Perplexity, Gemini, Grok, Google AI Mode, and Microsoft Copilot simultaneously, in parallel, with all data stored locally in your browser via IndexedDB, so no external database and no vendor lock-in are required.\n\n**OneGlanse** \u2014 It is open source, MIT licensed, and costs nothing to run on your own machine or VPS. Rather than calling model APIs, it opens the actual ChatGPT, Gemini, Perplexity, Claude, and AI Overview interfaces in a real browser, the same way a user would, and captures exactly what gets rendered: the full response, inline citations, recommended sources, and how your brand is positioned relative to competitors. It then uses your own OpenAI/Anthropic key to extract GEO scores, sentiment, visibility, rank position, competitor mentions, citation sources, and the AI perception breakdown.\n\n## Other GitHub-based developer tools\n\nAccording to a roundup of open-source AEO/GEO repositories, GetCito supports broad AEO and GEO experimentation, geo-aeo-tracker offers local-first monitoring, geo-optimizer-skill enables citation testing through Python and MCP, and AEOrank supports lightweight CI/CD checks. There are also lightweight MCP-server style tools built to track brand citations across ChatGPT, Claude, Perplexity, Gemini & Google AI Overviews, self-hostable on Cloudflare Workers, and a separate skill package that can check robots.txt, llms.txt, schema, citation signals, competitor gaps, and buyer prompts to gauge AI visibility readiness.\n\n## Zero-cost manual/DIY method (no tooling needed)\n\nIf you'd rather not stand up any software, teams can do this manually: query ChatGPT, Perplexity, and Google with your top 10-20 audience questions and log whether your brand appears, at what position, and which competitors are cited. It also helps to check server logs, since monitoring GPTBot, ClaudeBot, and PerplexityBot in your server logs reveals which pages AI platforms are actively indexing for citation. Specifically, when using ChatGPT's Search mode, if your site is cited, ChatGPT displays a URL reference alongside the answer.\n\n## Practical notes for a B2B marketing team\n\n- Most of the open-source options (Ansvisor, GEO/AEO Tracker, OneGlanse) require you to \"bring your own key\" for the underlying AI model or scraping API (e.g., Bright Data, OpenRouter, OpenAI, Anthropic), so there's a small usage cost even though the software itself is free.\n- Self-hosting gives you full data ownership \u2014 existing paid tools charge $200\u2013$500+/month, lock you into closed ecosystems, and store your data on their servers, which is the exact gap these open-source projects target.\n- Since non-determinism means the same prompt can produce different citations on different runs, any tool (open-source or not) should sample repeatedly rather than checking once.\n\nIf your team wants zero infrastructure to manage, Ansvisor and the Bright Data GEO/AEO Tracker are probably the most actively maintained and full-featured starting points; if you just want a quick gut-check with no setup, the manual ChatGPT/Perplexity query + server-log method costs nothing at all.", "refusal": null, "annotations": [ { "type": "url_citation", "url_citation": { "url": "https://www.ansvisor.com/blog/best-ai-search-citation-tracking-tools-for-websites-tested-compared", "title": "Best AI Search Citation Tracking Tools for Websites - Tested & Compared", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.ansvisor.com/blog/best-ai-search-citation-tracking-tools-for-websites-tested-compared", "title": "Best AI Search Citation Tracking Tools for Websites - Tested & Compared", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://github.com/aeohub/ansvisor", "title": "GitHub - aeohub/ansvisor: Open-source Answer Engine Optimization platform \u2014 track how ChatGPT, Google AI Overview, Google AI Mode, Perplexity, Claude, and Copilot mention your brand.", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://github.com/ansvisor/ansvisor", "title": "GitHub - ansvisor/ansvisor: Open-source AI Search Intelligence Platform \u2014 track, analyze, and improve AI visibility, citations, prompts, competitors, and content opportunities across ChatGPT, Claude, Gemini, Google AI Overviews, Google AI Mode, Perplexity, Grok, and Microsoft Copilot.", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://github.com/ansvisor/ansvisor", "title": "GitHub - ansvisor/ansvisor: Open-source AI Search Intelligence Platform \u2014 track, analyze, and improve AI visibility, citations, prompts, competitors, and content opportunities across ChatGPT, Claude, Gemini, Google AI Overviews, Google AI Mode, Perplexity, Grok, and Microsoft Copilot.", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://brightdata.com/blog/ai/ai-visibility-tracker", "title": "Open-Source AI Visibility Tracker With Bright Data", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://brightdata.com/blog/ai/ai-visibility-tracker", "title": "Open-Source AI Visibility Tracker With Bright Data", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://brightdata.com/blog/ai/ai-visibility-tracker", "title": "Open-Source AI Visibility Tracker With Bright Data", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://brightdata.com/blog/ai/ai-visibility-tracker", "title": "Open-Source AI Visibility Tracker With Bright Data", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://brightdata.com/blog/ai/ai-visibility-tracker", "title": "Open-Source AI Visibility Tracker With Bright Data", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://github.com/aryamantodkar/oneglanse", "title": "GitHub - aryamantodkar/oneglanse: The only free, open-source GEO tracker. Monitor how your brand appears in ChatGPT, Gemini, Perplexity, Claude, and AI Overview. \u00b7 GitHub", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://github.com/aryamantodkar/oneglanse", "title": "GitHub - aryamantodkar/oneglanse: The only free, open-source GEO tracker. Monitor how your brand appears in ChatGPT, Gemini, Perplexity, Claude, and AI Overview. \u00b7 GitHub", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://github.com/aryamantodkar/oneglanse", "title": "GitHub - aryamantodkar/oneglanse: The only free, open-source GEO tracker. Monitor how your brand appears in ChatGPT, Gemini, Perplexity, Claude, and AI Overview. \u00b7 GitHub", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.ansvisor.com/blog/top-github-repositories-for-aeo-and-geo", "title": "Top GitHub Repositories for AEO and GEO - Ansvisor", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://github.com/topics/ai-visibility", "title": "ai-visibility \u00b7 GitHub Topics \u00b7 GitHub", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://github.com/topics/ai-visibility", "title": "ai-visibility \u00b7 GitHub Topics \u00b7 GitHub", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.stackmatix.com/blog/ai-citation-tracking-tools", "title": "AI Citation Tracking Tools: Monitor Your Brand (2026)", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.stackmatix.com/blog/ai-citation-tracking-tools", "title": "AI Citation Tracking Tools: Monitor Your Brand (2026)", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://www.stackmatix.com/blog/ai-citation-tracking-tools", "title": "AI Citation Tracking Tools: Monitor Your Brand (2026)", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://github.com/scailetech/tech-geo-aeo-tracker", "title": "GitHub - scailetech/tech-geo-aeo-tracker: Open-source, local-first AI visibility intelligence dashboard. Track your brand across 6 AI models. \u00b7 GitHub", "start_index": 0, "end_index": 0 } }, { "type": "url_citation", "url_citation": { "url": "https://wrodium.com/blogs/top-8-ai-citation-tracking-tools-in-2026", "title": "Top 8 AI Citation Tracking Tools in 2026", "start_index": 0, "end_index": 0 } } ] } } ] }, "error": "EvidenceError" } ] }, { "engine": "gemini", "rows": [ { "question_id": "measure-problem", "pair_id": "measure-problem", "repeat": 1, "arm": null, "stage": "problem", "prompt": "How can a B2B marketing team tell whether AI answers name its brand?", "status": "error", "requested_model": "google/gemini-2.5-flash:online", "provider": "openrouter", "observed_at": "2026-09-10T15:14:34.009494+00:00", "error": "HTTP 404" }, { "question_id": "measure-category", "pair_id": "measure-category", "repeat": 1, "arm": null, "stage": "category", "prompt": "What kinds of tools can a B2B marketing team use to track citations in AI answers?", "status": "error", "requested_model": "google/gemini-2.5-flash:online", "provider": "openrouter", "observed_at": "2026-09-10T15:14:34.105216+00:00", "error": "HTTP 404" }, { "question_id": "measure-decision", "pair_id": "measure-decision", "repeat": 1, "arm": null, "stage": "decision", "prompt": "Which free, open-source tools can a B2B marketing team use to track whether AI answers cite its website?", "status": "error", "requested_model": "google/gemini-2.5-flash:online", "provider": "openrouter", "observed_at": "2026-09-10T15:14:34.176054+00:00", "error": "HTTP 404" } ] } ], "publication": { "approved_subject": "Skill Trade itself", "scope": "All 12 attempts from the first three-question run. One observation per question per model; no reruns are pooled into this sample.", "limits": "Only OpenAI completed 3 scorable answers. Other routes are unavailable, not zero. These unbranded questions were written before observing answers. This is a small API demonstration, not a market benchmark, before/after comparison or consumer-app measurement.", "raw_export": "Response projections retain returned model, completion status, full generated answer and URL annotations. Provider request identifiers, usage metadata, reasoning and retrieved source excerpts are omitted. Full raw responses remain private.", "original_run_sha256": "065a6b429dbfa654de56ea6833cbae692fbdc10eaadadde14082348ccd674565", "parser_version": 2, "replay": "Successful rows were reparsed from preserved responses; no new requests were made for this export." } }