Disclosure: SprueSentry is our own Labs site, not a customer story. Some offerings do not show in AI answers because assistants had nothing useful to cite and our pipeline briefly treated UI crumbs as products. We fixed the live-page verify requirement and the offering noise. We did not buy a mention rate.
Every number below comes from the public Labs feed or dated product receipts. If I cannot verify a figure, I do not invent one.
What "doesn't show" means here
SprueSentry is not on Shopify's Global Catalog. It is a generic Next.js catalog of kit pages and guides. When we say products do not show, we mean answer engines do not mention SprueSentry for the buyer questions we monitor.
Dated receipt from the Labs scoreboard (generated July 26, 2026; latest monitor runs July 18, 2026):
- Answer presence: 0 mentioned runs out of 47 total
- By engine: OpenAI 0/24, Google 0/23
- Recommended fixes: 14
- Published fixes: 1
- Verified fixes: 1
That is the observed floor. Measured since then, not necessarily because of any single change.
The actual requirement that failed
1. Artifact present (the page had to be live)
On July 17, 2026 we applied the recommended comparison fix: publish a "Which 40K boxes would you recommend?" guide versus Games Workshop framing, at https://spruesentry.com/guides/which-40k-boxes.
The product verify ladder recorded:
- 13:38 UTC: verify
artifact-present→ failed - 14:03 UTC: verify
artifact-present→ passed
The requirement is boring and absolute: the artifact you claim to have published must be reachable for the recrawl. A draft, a soft 404, or a still-deploying route fails the same way a Shopify product fails Catalog eligibility when it has no reachable product URL. We fixed the live URL, re-ran verify, and the check passed. Verified-live timestamp: July 17, 2026, 14:03 UTC.
2. Offerings must be real products, not UI crumbs
Earlier crawl work on SprueSentry pulled junk "offerings" such as numeric UI crumbs templated into questions like "Which 18 · spruesentry would you recommend?" Those are not products. They are extraction failures. They produce monitor questions no shopper asks and no honest page can answer.
The fix is upstream of content: filter offerings so the question set is grounded in real boxes, armies, and guides. Product research and the answer-monitor quality rails document that failure class. We treat it the same way a Shopify merchant should treat a SKU-titled CSV import: the catalog entry is wrong before any ranking conversation starts.
What we did about it
One variable at a time, logged before outcomes were known:
- Publish the comparison guide AgentReady recommended (content-item artifact).
- Confirm it is live via verify-by-recrawl (failed, then passed the same day).
- Keep answer monitoring on a Labs cadence and read the next run without rewriting history.
We did not claim the guide would produce mentions. The Labs fix record still shows after: null / no post-fix mention lift attached. Honesty rail on the record: "Measured since then, not necessarily because of it."
What we observed after
Next monitor batch after the verified publish (July 18, 2026): still 0 mentions across the OpenAI and Google runs in that window. Total Labs tally remained 0/47 as of the July 26 feed snapshot.
That is the observed-after result. It is not a failure of the article to ship. It is evidence that one verified page does not equal assistant citations on a ten-day-old Labs site. Anyone selling you a guaranteed mention lift from a single publish is selling you a story we refuse to tell about our own property.
The Shopify parallel (this is the bookmarkable part)
If you sell on Shopify, swap the substrate and keep the method:
| SprueSentry failure | Shopify Catalog parallel | Where to check |
|---|---|---|
| Page not reachable for verify | Product not published to Online Store / Hydrogen / Headless, or no reachable URL | Admin → product → sales channels; open the product URL logged out |
| Junk offerings in the question set | Empty titles, SKU-as-title imports, $0 prices, missing images | Eligibility checklist |
| 0 mentions on monitored questions | Absorbed but never surfaces on category search | Absorption check |
Organic ChatGPT product surfacing for Shopify merchants is Shopify Catalog. Agentic Storefronts have been on by default for eligible stores since roughly March 2026. OpenAI's Instant Checkout was effectively retired in March 2026, so do not plan around in-chat checkout. Plan around catalog eligibility, clean data, and dated measurement.
How to run the same loop without installing anything
- Pick five buyer questions a stranger would ask (not your brand name).
- Ask ChatGPT and one other assistant. Record presence, who appears, and the date.
- For each missing product, walk the six eligibility rules by hand in Shopify admin.
- Fix one requirement. Wait for Catalog re-ingestion (hours to days; no published SLA).
- Re-ask the same five questions. Log the diff.
AgentReady automates the verify and re-check half of that loop. It does not replace the eligibility walk. If deleting every AgentReady sentence above still leaves you a usable checklist, the article did its job.
For the build story and Labs disclosure in full, see how we built SprueSentry for the AI catalog era. To run the free merchant check, use the Catalog Readiness checker, then the /shopify lander for what we verify on install.
Comments
Every comment here comes from a verified email. Write yours, confirm from your inbox, and it's live.
Loading comments…