<- all tokdocs
2026.10.24@jamout.ai

The video credits 13,000 citations to ChatGPT and Claude; the dashboard on screen is Microsoft's AI Performance report, which Microsoft documents as counting Copilot and Bing citations

Marketing
Watch on TikTok

View on TikTok ->

*This is a 56.166667 s vertical clip, 1080x1920, QuickTime/MP4 container (mov,mp4,m4a,3gp,3g2,mj2), HEVC Main video in an hvc1 tag at a constant 30/1 fps across 1,685 frames at 762,017 bps, with HE-AACv2 audio at 44,100 Hz stereo and 64,213 bps, an overall 833,045 bps, a 5,848,677 byte file, format id bytevc1_1080p_833041-1, and an audio stream measuring 56.143991 s. It was posted at epoch 1792848489, which is 2026-10-24 13:28:09 UTC, by @jamout.ai (uploader ID 6779549123316057093) under the nickname "Jam | AI & Marketing", with the audio credited as "original sound" by "Jam | AI, Marketing" and no licensed music track. The description runs 78 characters. At the capture time recorded by the metadata.json file mtime, 2026-10-24 18:04:21 local, the counters read 5,287 views, 235 likes, 10 comments and 157 reposts. I read 17 of the 28 extracted frames (001, 003, 004, 005, 006, 007, 009, 011, 013, 014, 016, 018, 020, 022, 024, 026, 028, which sit at 0, 4, 6, 8, 10, 12, 16, 20, 24, 26, 30, 34, 38, 42, 46, 50 and 54 seconds) against a 23-segment, 190-word transcript. A white sticker reading "Entity Map: you need one" sits at the top of every frame I read. Frame 001 (0 s) is a talking-head shot in front of a pebble-pattern wall with the burned-in caption "ChatGPT can't stop talking about me". Frames 003 and 004 (4 and 6 s) overlay a screenshot card headed "AI Performance", subhead "Citation sources: Microsoft Copilots and Partners", with tabs "Citations & Pages" and "Topic Portfolio BETA", a range row "7 D / 30 D / 3 M / 6 M / Custom / Compare" with "30 D" selected, and tiles reading "Total Citations 13.2K" and "Avg. Cited Pages 17". Frames 005 through 007 and 013 through 014 (8 to 12 s and 24 to 26 s) are a phone-held shot of a laptop showing a page headed "Jam AI EntityMap" with the lines "readable version: entitymap.json. Format: EntityMap v1.0." and "Updated 2026-10-23 · 24 entities · self-declared", above a two-column entity list including "Jam AI Organization", "Robyn "Jam" Anderson Person", "Club Claude Service", "Zero to Claude Methodology", "Jammer SoftwareProduct" and "Six brand files ProprietaryTerm", followed by entity cards with RELATIONS rows such as "OFFERS Club Claude" and "TARGETS AI marketing for small business (declared)" plus EVIDENCE blockquotes. Frames 009 and 011 (16 and 20 s) show the raw file in a browser tab, with visible keys "1.0", "https://entity...", "2026-10-23T17:27:22Z", "core", "self-declared", "entityId": "e_001", "Organization", "alternateName": "jamout.ai", a chunkId of c_001 carrying sourceUrl, pageTitle, publisher, retrieved and contentType, and a relations array of {"predicate": "OFFERS", "targetId": "e_003", "targetName": "Club Claude"} objects. Frames 016 through 024 (30 to 46 s) walk a web tool: a card numbered "07" headed "Entity Map Builder" with the strapline "Build the file that tells ChatGPT, Claude and", a form reading "Jammer reads your website, picks out what yo... entitymap.org) with quotes copied word for wo..." with "Your website " filled in as https://friendassembly.com/, a red "Map my website" button annotated "Takes about 60 to 90 sec...", a results panel headed "Your Entity Map" reading "I read 6 pages of friendassembly.com and mapped...", a "STEP 1 · CHECK IT" checklist with rows "Friend Assembly BUSINESS", "Stanley Marketplace PLACE" and "Wheel Throwing TOPIC", a "STEP 2 · DOWNLOAD IT" block with the checkbox "I checked these and they're accurate", the green line "14 things you're known for, passing every EntityMap v1.0 check", buttons reading "Download entitymap.json" and "Download entitymap.html", and a "STEP 3 · PUT IT ON YOUR WEBSITE" dropdown listing ten platforms from "WordPress" to "Vercel, Netlify or custom code". Frames 026 and 028 (50 and 54 s) return to the talking head, closing on "Join me in Club Jam. Link in bio". No frame I read shows a ChatGPT window, a Claude window, a Google Search Console property, a schema.org validator, a Rich Results Test, or a before-and-after citation chart tied to the entity map being published. The spoken argument is that a large citation count follows from having a good entity map, and that a builder tool inside the creator's paid community generates that file for members. The transcript's last segment runs 54.38 s to 55.90 s, inside both the 56.143991 s audio stream and the 56.166667 s container, so there is no Whisper overshoot and no fabricated tail text.

The 13,000 figure comes from a Microsoft dashboard, and Microsoft does not say ChatGPT or Claude are in it

At 0.00 s to 5.00 s the video says "ChatGPT and Quad have cited my website over 13,000 times this month". The screenshot that carries that number appears at frames 003 and 004 (4 and 6 s), and it labels itself. The card is headed "AI Performance", its subhead reads "Citation sources: Microsoft Copilots and Partners", the selected range pill is "30 D", and the tile reads "Total Citations 13.2K". That is the AI Performance report in Bing Webmaster Tools. Microsoft announced it on 10 February 2026, describing the coverage as "Microsoft Copilot, AI-generated summaries in Bing, and select partner integrations" and defining the metric as the number "of citations that are displayed as sources in AI-generated answers during the selected time frame". Microsoft names no partner integrations in that announcement, and neither ChatGPT nor Claude is mentioned in it.

The video names two assistants Microsoft does not claim to measure and omits the two surfaces Microsoft does claim to measure. OpenAI's own crawler documentation states that "OAI-SearchBot is used to surface websites in search results in ChatGPT's search features", which describes an OpenAI-operated index for ChatGPT search rather than a Microsoft one. I tried to reach Anthropic's published subprocessor list to establish which provider backs Claude's web search and could not load it, so I make no claim about Claude's retrieval stack. What stands is checkable: the only document defining the number on screen attributes it to Microsoft Copilot, Bing AI summaries, and partners Microsoft declines to name.

The file the tool produces is EntityMap v1.0, and it is not schema.org markup

At 40.50 s to 44.44 s the video says "it actually creates that JSON schema markup file for you". In SEO usage, "schema markup" means schema.org vocabulary, and Google's structured data introduction states that "Most Search structured data uses schema.org vocabulary" delivered as JSON-LD "embedded in a <script> tag in the <head> and <body> elements of an HTML page". The artifact the tool produces carries a different name, and the tool says so on screen. Frame 022 (42 s) labels the download buttons "Download entitymap.json" and "Download entitymap.html", and the green validation line above them reads "passing every EntityMap v1.0 check".

I fetched the live file the video is demonstrating. https://jamout.ai/entitymap.json returns 200 with content-type: application/json; charset=utf-8 and a 32,499 byte body whose top-level keys are version, schema, publisher, generated, profile, verificationStatus and entities. The schema value is https://entitymap.org/spec/v1.0, matching the "Format: EntityMap v1.0" line in frames 005 and 006. The file parses cleanly under both Python's json module and Node's JSON.parse, and it carries 24 entities, matching the count rendered on the page in frame 005. It does not contain an @context key. Its first entity object has the keys entityId, @type, name, description, alternateName, hasChunks and relations, which is the structure legible in frames 009 and 011.

The @type values are where the schema.org framing breaks. Across the 24 entities the file uses Organization, Person, Service, Event, SoftwareProduct, Methodology, ProprietaryTerm, Concept, jam:DigitalProduct, jam:Keynote and jam:Podcast. I requested each bare name against schema.org directly. Organization, Person, Service and Event return HTTP 200. SoftwareProduct, Methodology, ProprietaryTerm and Concept return HTTP 404, so schema.org does not define them, and the jam: prefixed values are the publisher's own namespace. The relations array uses a predicate, targetId, targetName shape and the evidence array uses chunkId, text, sourceUrl, pageTitle, publisher, retrieved and contentType, none of which exist in schema.org's vocabulary.

Using @type without @context makes the file look like JSON-LD at a glance. The JSON-LD 1.1 specification explains that a context "is used to map terms to IRIs", and that a plain JSON document can be read as JSON-LD by "referencing a context via an HTTP Link Header". The only link header on https://jamout.ai/entitymap.json is </.well-known/ai-catalog.json>; rel="ai-catalog", which is not a JSON-LD context relation, and the content type is application/json rather than application/ld+json. As served, a JSON-LD processor has no context to resolve Organization or SoftwareProduct against. The file is structured JSON on its own terms, which is what EntityMap v1.0 asks for, and it is not the schema.org JSON-LD that the phrase "JSON schema markup" points at.

EntityMap is a real proposed standard with an identifiable author, and no vendor has endorsed it

The video never names the format out loud, which makes the on-screen text the better record. EntityMap is a documented open standard published at entitymap.org, described there as "a proposed open standard for publishing a structured, entity-first index of website knowledge for AI systems". The site credits the work as "Initiated by Fred Laurent with the support of Dixon Jones" and associates it with Waikay and InLinks Optimization Ltd. The v1.0 specification is dated 2026-04-07, carries a status of Stable and a CC BY 4.0 licence, and frames itself against sitemaps rather than against schema.org: "Where sitemap.xml tells crawlers what pages exist, entitymap.json tells AI systems what a site knows."

Two omissions matter, because the video's framing implies otherwise. The specification makes no claim of adoption by any AI vendor or search engine, and the site lists exactly one implementation, the Waikay EntityMap Generator. Its certification registry is not live, with the specification stating that "The live registry launches Q3 2026". The file the builder produces conforms to a published standard, and that standard does not assert that any assistant currently reads it.

Google's documentation says no file of this kind is needed to appear in AI answers

The video's premise is that a machine-readable description of a business is what gets that business into AI answers. Google addresses that premise directly in its AI features documentation for site owners: "There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary." The same page continues: "You don't need to create new machine readable files, AI text files, or markup to appear in these features. There's also no special schema.org structured data that you need to add."

That statement scopes to Google's AI Overviews and AI Mode, so it does not settle what Microsoft Copilot or any assistant weighs. It does rule out the strongest version of the claim for the largest AI answer surface with published guidance. Neither Google's position nor OpenAI's crawler documentation forecloses that an entity page helps, since a well-structured HTML page at /entitymap is still a crawlable page that states what a business does. The gap is between "a clear page about your business can be read and cited" and "this specific JSON file is why ChatGPT cites me".

The causal claim is asserted at 5.00 s and never demonstrated in any frame

At 5.00 s to 7.60 s the video says "and it's because I have a super good entity map". That is a causal claim about the 13.2K number, and the frames do not support it. The entity map was generated on 2026-10-23 at 17:27:22 Z, which is the generated timestamp in the live JSON file and the "Updated 2026-10-23" line in frame 005. The video was posted at 2026-10-24 13:28:09 UTC, about twenty hours later. The dashboard tile reads "30 D", so the 13.2K count covers a window almost entirely before the entity map existed in its current form. The dashboard also reports "Avg. Cited Pages 17", so the citations spread across seventeen pages per day on average, a site-wide pattern rather than one page's effect.

No frame I read shows a comparison view, and the "Compare" pill visible in frames 003 and 004 is unselected. The site also publishes https://jamout.ai/llms.txt, which opens "Jam AI is a Denver-based AI marketing company for everyday, non-technical small business owners. Founded in 2023 by Robyn "Jam" Anderson", so there are at least two machine-readable artifacts on the domain plus ordinary indexed pages, and nothing on screen separates their contributions. What the frames establish is a site with a high citation count in Microsoft's report and a recently generated entity map, with no evidence presented that links them.

Whisper rendered "Claude" as "Quad", and the frames settle it

The transcript at 0.00 s reads "ChatGPT and Quad have cited my website". "Quad" is not a product. The burned-in screen text resolves it: frames 005, 006 and 013 list "Club Claude Service", "Zero to Claude Methodology" and "Claude Soft..." in the entity list, frame 014 describes the founder as "an Official Claude Partner", frame 016 shows the builder strapline "Build the file that tells ChatGPT, Claude and", and the live entitymap.json contains "targetName": "Club Claude". The spoken word is Claude. I quote the transcript's literal "Quad" only here, so the misreading is on the record rather than silently smoothed over.

Key Takeaways

  • Correction: At 0.00 s the video says "ChatGPT and Quad have cited my website over 13,000 times this month". The dashboard supplying that number, visible in frames 003 and 004, is headed "AI Performance" with the subhead "Citation sources: Microsoft Copilots and Partners". Microsoft's 10 February 2026 announcement scopes that report to "Microsoft Copilot, AI-generated summaries in Bing, and select partner integrations" and names no partners, and mentions neither ChatGPT nor Claude.
  • Correction: At 40.50 s the video says the tool "creates that JSON schema markup file for you". The buttons in frame 022 read "Download entitymap.json" and "Download entitymap.html", and the live file's schema value is https://entitymap.org/spec/v1.0. Its @type values include SoftwareProduct, Methodology, ProprietaryTerm and Concept, all of which return HTTP 404 on schema.org, plus three jam: prefixed custom types.
  • Correction: Whisper transcribed "Claude" as "Quad" at 0.00 s. The frames show "Club Claude", "Zero to Claude" and the builder strapline "tells ChatGPT, Claude and", and the live JSON contains "targetName": "Club Claude".
  • Partial correction: At 5.00 s the video treats the entity map as the cause of the citation count. The file's generated timestamp is 2026-10-23T17:27:22Z and the dashboard range is "30 D", so most of the counted window predates the file. No before-and-after view appears in any frame I read.
  • Clarification: Google's AI features documentation states "You don't need to create new machine readable files, AI text files, or markup to appear in these features. There's also no special schema.org structured data that you need to add." That scopes to Google's AI surfaces and does not speak for Microsoft Copilot or any assistant.
  • Clarification: The served file uses @type without @context, and its only HTTP link header is rel="ai-catalog", not a JSON-LD context relation. Under JSON-LD 1.1 the file as served has no context to resolve its terms against.
  • Verified: The entity map is real and live. https://jamout.ai/entitymap renders with the header "Jam AI EntityMap" and the line "Updated 2026-10-23 · 24 entities · self-declared", and https://jamout.ai/entitymap.json returns a 32,499 byte file that parses cleanly and contains exactly 24 entities. Both match the frames.
  • Verified: EntityMap is a documented open standard. entitymap.org publishes a v1.0 specification dated 2026-04-07, status Stable, CC BY 4.0, credited as "Initiated by Fred Laurent with the support of Dixon Jones". It claims no vendor adoption, lists one reference implementation, and states that its certification registry "launches Q3 2026".
  • Context: The demo in frames 016 through 024 runs against https://friendassembly.com/, a Denver ceramics studio, reporting "I read 6 pages" and "14 things you're known for". The deployment dropdown in frame 024 lists ten platforms, broader than the "Squarespace or WordPress" named at 47.20 s.
  • Unverified: Which "select partner integrations" Microsoft includes. Microsoft does not disclose them, so no claim about whether a third-party assistant feeds that 13.2K figure can be checked.
  • Unverified: Which provider backs Claude's web search. I could not load Anthropic's published subprocessor list, so I make no claim.

Resources

Published October 24, 2026. Writeup generated from a favorited TikTok.