Practice sample · 2-minute read
AI knows Cronhooks exists — but doesn't recommend it
Cronhooks schedules one-time and recurring webhooks. When we named it in a prompt, every assistant we tested described it. When a buyer asked a generic question, ChatGPT, Claude and Perplexity never named it (0 of 9 answers).
Five buyer prompts, four assistants
Position = where Cronhooks first appeared among the tools named (early / mid / late), or not named. One signed-in run per prompt per assistant.
| # | Buyer prompt | ChatGPT | Claude | Perplexity | Grok |
|---|---|---|---|---|---|
| P1 | “What's the best cron job scheduling service for triggering webhooks?” | Not named | Not named | Not named | Mid |
| P2 | “What are some good alternatives to EasyCron?” | Not named | Not named | Not named | Not named |
| P3 | “How can I schedule HTTP requests on a cron without running my own server?” | Not named | Not named | Not named | Mid Passing mention |
| P4 | “cron-job.org vs Cronhooks — which one should I use for scheduled webhooks?” (branded) | Described | Described | Described | Described |
| P5 | “I need to send a webhook at a specific future date and time (like a reminder 3 days after signup) in the user's timezone — is there an API service that does this?” | Late Hedged: “after verifying its API” |
Not named | Not named | Early |
What the table says
- Generic prompts (P1–P3): ChatGPT, Claude and Perplexity named Cronhooks in 0 of 9 answers. Grok named it mid-list on P1 and P3.
- Branded prompt (P4): all 4 assistants knew it — one-time and recurring schedules, timezones, signed requests, failure alerts. Awareness isn't the problem.
- Its own sweet spot (P5): a future-dated webhook in the user's timezone is exactly what the homepage describes, yet only Grok put it early.
- Who got named instead, often: Upstash QStash, cron-job.org, Google Cloud Scheduler, EventBridge Scheduler, Posthook, Trigger.dev, Inngest.
What an assistant finds on cronhooks.io
Direct checks with curl, Oct 9, 2026 (IST). These don't decide ranking on their own, but they shape what an assistant can confidently say.
1. No comparison pages, no FAQ
No “vs” or alternatives pages in the sitemap or navigation, and the site doesn't name a competitor anywhere. There is no FAQ page or FAQ section. So when a buyer asks “alternatives to EasyCron?”, the pages that get quoted are other people's.
2. /llms.txt tells a different story than the homepage
Having an /llms.txt at all puts Cronhooks ahead of many peers. But it gives most of its weight to cFlows (AI workflow automation), which the homepage doesn't mention; the homepage leads with “Webhooks, fired on time. Every time.” Two of its anchor links (#cflows, #cRequest) don't resolve to a section.
3. The About page is still the site template
/about returns 200 and is listed in the sitemap, but it contains the starter template's placeholder text and two Lorem ipsum paragraphs. It isn't linked from the navigation, but crawlers can still reach it via the sitemap.
4. Thin machine-readable signals
The JSON-LD is a generic Organization block only (no SoftwareApplication, no pricing Offer, no FAQPage). Pricing ($0 / $5 / from $12 a month) exists only as a homepage section, with no standalone page. robots.txt doesn't point to the sitemap.
What we'd fix first
- Publish 2–3 honest comparison pages and a short FAQ. Start with the names assistants already use: cron-job.org, Upstash QStash, EasyCron. Say plainly when Cronhooks is the better pick (one-off, per-user, timezone-aware webhooks) and when it isn't. Add one FAQ answer for the “webhook at a future date in the user's timezone” question.
- Make
/llms.txtlead with webhook scheduling. Match the homepage's one-line positioning, keep cFlows as a secondary section, and fix the two anchors. - Replace the About page and add basic structure. A few real lines about who runs Cronhooks;
SoftwareApplication+OfferJSON-LD with the published pricing; aSitemap:line inrobots.txt.
How this was measured: 5 buyer prompts, each pasted once into a new chat in signed-in ChatGPT, Claude, Perplexity and Grok on Oct 9, 2026 (IST) — 1 run per prompt per assistant, default settings. Chat memory and account history were not controlled, and answers vary run to run, so treat each cell as a single dated snapshot, not a rate. Gemini was skipped because it wasn't signed in on our browser. Site checks were made with curl the same day.
Want this for your product?
Send your product URL and your top 2 competitors. I'll run your buyer prompts and send back a short write-up like this one: who gets named, and the first three things I'd fix. No cost, no obligation.
Longer practice example: Checkly sample audit. CiteSprint ethics: no citation guarantees · no fake reviews · no spam · no invented chat transcripts.