GEO Lab #03 · S3-GROK-M2-R2 · Original response
This page presents one selected response from GEO Lab #03. It adds a smaller reading and citation entry point; it does not replace the full archive or collect a new answer.
The OpenAI arm uses the Responses API with web search and does not represent consumer ChatGPT Search. Grok records identify their actual consumer-interface mode and available model label.
Recorded conditions
Original prompt
As a bootstrapped SaaS builder, what's a free, repeatable workflow for checking whether my product appears in ChatGPT and Grok recommendations?
The agreed prompt is reproduced exactly. Any recorded difference in the actual submitted text or search UI is retained in the raw record and full appendix.
Original answer
A free, repeatable workflow is a fixed set of buyer-intent prompts + manual sampling in fresh sessions + a simple spreadsheet. This is the standard DIY approach used by bootstrapped founders and small teams for ChatGPT and Grok (and other engines). Paid trackers exist, but they are unnecessary for a baseline and trend view. 1. Build a fixed prompt set (do this once) Create 8–15 prompts that real buyers would type before they know your brand name. Freeze the exact wording so you can compare over time. Core types: Category discovery: “What’s the best [your category] for [specific use case / company size / audience]?” Alternatives: “Best alternatives to [leading competitor]” Comparisons: “[Competitor A] vs [Competitor B] for [use case]” Problem / job-to-be-done: “How do I [the outcome your product delivers] without [common pain]?” Pull language from customer calls, support tickets, Google’s “People also ask,” Reddit, or your own Search Console queries. Avoid pure branded questions (“Is [your product] good?”) for discovery checks—they inflate results. Keep branded prompts only for accuracy checks (pricing, features, downsides). 2. Run the checks cleanly ChatGPT (chatgpt.com): Use a Temporary Chat and choose the unpersonalized option (or log out / use a private window). This reduces memory and personalization bias. Prefer sessions with web search enabled when available. Run each prompt 2–3 times in fresh chats. Grok (grok.com or via X): Start a fresh conversation each time. Grok draws heavily from X posts in addition to the web, so its recommendations can differ from ChatGPT. Same 2–3 runs per prompt. For every answer, note: Did your product appear? (Y/N) Position / order if it was in a list Competitors named Sources cited (especially useful) Brief framing (how it described you or why it preferred someone else) 3. Log it in a spreadsheet One row per run. Columns that matter: | Date | Engine | Prompt | Mentioned (Y/N) | Position | Competitors | Citations | Notes | At the end of each cycle, calculate simple metrics: Mention rate (how often you appear on unbranded prompts) Share of voice vs the 2–3 competitors that keep showing up Citation rate (how often your site or a trusted third-party page about you is linked) A Google Sheet or Notion table is enough. No special tools required. 4. Cadence Run a full baseline now. Repeat monthly (or every 2 weeks on your 5–6 highest-priority prompts). Extra spot-checks after major content launches, new reviews (G2/Capterra/etc.), or big X activity. Time cost is typically 20–40 minutes per cycle once the prompts are set. Answers are non-deterministic—the same prompt can produce different shortlists across sessions, models, and dates—so the trend across multiple runs matters more than any single screenshot. Practical notes for bootstrapped SaaS Free one-shot checkers (Ahrefs Brand Radar free preview, various “AI visibility” free tools) can give a quick extra data point, but they don’t replace your controlled, repeated prompt set. Ensure your site is crawlable (don’t block major AI crawlers in robots.txt). For Grok specifically, an authentic presence on X helps because of its heavier weighting of X content. This measures whether you appear in recommendations. Improving it usually means earning consistent third-party coverage (listicles, reviews, comparisons) plus clear, accurate pages on your own site—not paid placements (OpenAI has stated product results are not ads). This workflow is free, fully under your control, and gives a clear before/after view of whether ChatGPT and Grok are recommending your product. Start with the prompt list and a spreadsheet this week; the rest is just consistent sampling.
Recorded sources
These are the source roles recorded in the public dataset. A retrieval or source-panel item is not automatically an answer-body citation. A URL supplied in the prompt and repeated by an answer does not prove content access. No destination has been repaired or independently validated here.
Retrieval-only / source-panel inventory (2)
- stork.ai/blog/does-chatgpt-recommend-your-productBrowsed link in search-activity panel; not an inline answer citation · www.stork.ai
- arahi.ai/blog/how-to-check-brand-visibility-in-chatgptBrowsed link in search-activity panel; not an inline answer citation · arahi.ai
Evidence and qualifications
This is the existing redacted public record. It has not been rewritten for this page. Earlier privacy edits and collection qualifications continue to apply. Editorial text is English; a few original captures retain non-English UI badges or source-panel date labels. An answer’s empirical claims are platform output, not verified findings of this archive.
Capture method
Complete rendered Grok article innerText; UI heading/duration and source badge removed. Exact inline hrefs extracted from scoped rendered article.
Record notes
[
"Source panel groups results by web-search query; 65 is the sum of reported result hits, not unique answer citations. Complete body has no inline citation anchors."
]
Existing public-text privacy policy
{
"body_changed": false,
"policy": "Only actual URL tracking, signed/session URL parameters and necessary identifying information are removed; original wording, including analytics parameter examples outside URLs, remains.",
"annotation_positions": "Original position indexes omitted because URL/privacy edits can change character locations."
}
Prompt SHA-256: 9d96cab49de6d93cb0c1eef3c1da1e5ab5cfdf02f4a703ce57ecf5911dc08219
Answer SHA-256: 2b89ab8acc0f5bec7a638b85acc08603eea85bb29c8e610bd85f8936f4f37fdc