How to Rank on ChatGPT and Claude | Generative Engine Optimization (GEO) Webinar

Summary

Digital Niche Agency ran a webinar on generative engine optimization, the practice of shaping a brand's presence so that ChatGPT and Claude retrieve it, cite it, and name it in answers. CEO Jason Fishman and Head of Content Khalil Doheny present a six-part system: building a prompt panel, taking a baseline reading, mapping the sources models actually pull from, restructuring pages into answer-ready form, building authority signals, and then measuring what moved.

The practical content sits in the details rather than the headline. The presenters argue that ChatGPT and Claude need to be tested as separate environments rather than treated as one AI channel, that a technical access test should come before any content rewrite, that pages work better when built around a complete decision including the conditions under which the product is a poor fit, and that Reddit threads and Google reviews carry weight out of proportion to their size in what models recommend. They also describe monitoring for answer drift week over week. The deck is published as a downloadable Google Slides file, and the agency positions the framework as the one it uses on its own marketing and on investor acquisition campaigns.

Demonstration by Digital Niche Agency, embedded from YouTube. AIOTruth did not reproduce the test.

Why it matters

The specific shift here is from optimizing a page for a ranking position to assembling an evidence trail a model can assemble an answer from. If the claim about separate testing environments holds, then a brand that checks its visibility in one assistant and assumes the result generalizes is measuring roughly half of its exposure, and any remediation it plans off that single reading is aimed at the wrong gaps. The emphasis on third-party surfaces has a similar consequence: a brand can control its own site completely and still lose the recommendation, because the material shaping the answer sits in review platforms and community threads it does not own and cannot rewrite. Building pages around complete decisions, including stated poor-fit conditions, points the same direction, since a model summarizing a purchase decision needs the disqualifying conditions to produce a useful answer, and a page that only sells cannot supply them. For anyone working on being found by AI systems, the operational takeaway is that the audit surface is wider than the website, the measurement has to be repeated per model and over time to catch drift, and access needs verifying before content work begins rather than after it fails to move anything.

Source

• docs.google.com/presentation/d/1Ev7K5YsQ22JtSWtnAA4KeyA43wQEjALBRGpgPMpfBV0/edit?usp=shari the item itself, published by Digital Niche Agency, created by Digital Niche Agency
• www.digitalnicheagency.com/ Reputable reporting
• youtube.com/watch?v=ClT7JBQFJGAwhere AIOTruth found it

What was checked

• read the item page (200)
• https://docs.google.com/presentation/d/1Ev7K5YsQ22JtSWtnAA4KeyA43wQEjALBRGpgPMpfBV0/edit?usp=sharing: does not mention the item, tier 8 (Named expert analysis, or a first-party claim about itself)
• https://docs.google.com/_/presentations/_/js/k=presentations.editor_js_prod.en.wFT7XqtB0Eo.es5.O/am=QIAIcABg/d=0/wt=0/rs=AB6fld0dpHt74ZD-VoXGwpTuxoN9pAnlSg/m=core: does not mention the item, tier 8 (Named expert analysis, or a first-party claim about itself)
• https://workspace.google.com/products/slides/: does not mention the item, tier 8 (Named expert analysis, or a first-party claim about itself)
• https://accounts.google.com/v3/signin/identifier?amp;continue=https://docs.google.com/presentation/d/1Ev7K5YsQ22JtSWtnAA4KeyA43wQEjALBRGpgPMpfBV0/edit?usp%3Dsharing&ec=GAZAmQI&followup=https://docs.google.com/presentation/d/1Ev7K5YsQ22JtSWtnAA4KeyA43wQEjALBRGpgPMpfBV0/edit?usp%3Dsharing&ltmpl=slides&osid=1&passive=1209600&service=wise&flowName=WebLiteSignIn&flowEntry=ServiceLogin&dsh=S1680800003:1786252903831397: does not mention the item, tier 8 (Named expert analysis, or a first-party claim about itself)
• https://www.digitalnicheagency.com/: supports the item, tier 9 (Reputable reporting)

Limitations

• Everything checked traces back to a single domain, so this is one party's account.
• No official documentation from the platform or vendor was found for this item.
• The demonstration was not independently reproduced by AIOTruth.

AIOTruth judgment

The item scored 7.7, carried by relevance of 10 and usefulness of 9. It sits directly on the question of how brands get retrieved and recommended by AI systems, and the six-part structure gives someone a sequence they can actually run rather than a set of principles. Evidence and originality both scored 5, and the Emerging status follows from that. Everything checked traces back to a single domain, the agency's own site, so this is one party's account of a method it sells. No platform or vendor documentation was found that corroborates any of the mechanics described, and AIOTruth did not reproduce the demonstration. The commercial results cited, including the capital raise and the single-ad investment figure, are the agency's own reporting with no external confirmation available. The claim about search traffic migration is presented without a source in the material. The framework components themselves largely restate approaches circulating across the GEO field rather than introducing new mechanics, which is what holds originality where it landed. The rating reflects a useful and well-organized practitioner account whose supporting evidence has not been checked outside its author.

Watch the demonstration Open the source

Found something wrong here? AIOTruth corrects material errors openly. Have something we should review? Submit a find.