Who sets your Auckland clinic's opening hours — you, or a machine's guess?

This week Google started testing a feature where AI reads a business description and sets the opening hours itself. We spent the morning of 25 September checking eight independent Auckland dental clinic websites: six of the eight do not publish their opening hours in machine-readable form. When the machines go looking, they find nothing — and guessing is what they'll do next.
What Google just started doing
On 24 September 2026, Search Engine Roundtable reported that Google Business Profiles is testing a feature to set business hours automatically — you describe your hours in text, Google's AI turns them into the structured panel that shows in Search and Maps (read it yourself, dated 24 Sep 2026).
Sounds convenient. And it is — if the machine can find your real hours somewhere reliable. If it can't, it still has to produce an answer. Assistants do not say "I don't know your clinic's hours." They say something.
This morning we looked at eight Auckland clinics
Between 09:00 and 09:45 IST on 25 September 2026, we fetched the homepages of eight independent Auckland dental clinics (found the way anyone finds them — a plain web search) and counted two things: does the page carry any structured data, and does it state opening hours in machine-readable schema.
| Clinic (by suburb) | Homepage loads | Any structured data | Machine-readable hours |
|---|---|---|---|
| Milford, North Shore | ✓ | ✓ (1 block) | ✗ |
| Drury, South Auckland | ✓ | ✓ (2 blocks) | ✗ |
| Parnell (independent) | ✓ | ✓ (1 block) | ✗ |
| New Lynn, West | ✓ | ✓ (2 blocks) | ✗ |
| Manukau (clinic A) | ✓ | ✓ (2 blocks) | ✗ |
| Manukau (clinic B) | ✓ | ✓ (2 blocks) | ✓ |
| Citywide network site | ✓ | ✓ (1 block) | ✗ |
| Parnell (network site) | ✓ | ✓ (4 blocks) | ✓ |
Six of eight: no machine-readable hours. We don't name names — the point is the pattern, not the clinics. Their websites show hours beautifully to humans ("Mon–Fri 8–5, Sat by appointment") and say nothing to machines. Raw data is kept in our working files and shared on request: hello@synapsein.com.
Test your own clinic in 60 seconds — honestly
1. Open your clinic's homepage.
2. Right-click → View page source (Ctrl+U / ⌘⌥U).
3. Ctrl+F and search for openingHours (case-sensitive).
4. Nothing found? Then to Google, Bing, ChatGPT, Perplexity — everyone — your hours are folklore, not fact.
Try the same search for ld+json to see whether your site carries any structured data at all. The 100%-certain check: run your homepage through Google's free Rich Results Test.
What the fix actually looks like
<script type="application/ld+json">
{ "@context":"https://schema.org", "@type":"Dentist",
"name":"Example Dental, Milford",
"telephone":"+64-9-000-0000",
"openingHoursSpecification":[
{ "@type":"OpeningHoursSpecification",
"dayOfWeek":["Monday","Tuesday","Wednesday","Thursday","Friday"],
"opens":"08:00","closes":"17:00" } ] }
</script>
Paste once, verify with Google's own free tester — and now your statement, not a language model's inference, is what the machines carry. The same pattern covers address, services, practitioners and appointment links.
Who's writing this (and why you should verify us, not trust us)
We're SynapseIN — a small research-and-build team based in India, working with clinics and practices on how machines represent them. We're 7½ hours behind New Zealand: email or WhatsApp whenever suits; calls book best 4–7pm NZ time. We keep a public daily ledger of 36 first-party sources — the Policy Radar — with hash-verified snapshots. Our field is machine readability: structured data, citation hygiene, the boring plumbing that decides whether ChatGPT recommends your clinic or invents it.
Questions this post keeps getting
Does Google really set business hours by itself now?
It's in testing, reported 24 September 2026: Google Business Profiles will take a plain-text description of your hours and structure them using AI. If your site doesn't supply verified hours, the AI works from whatever it finds — or doesn't.
My site shows opening hours to visitors. Isn't that enough?
For humans, yes. For machines, no — unless those hours also exist in structured form (the openingHours schema). Our Auckland sample: 6 of 8 clinics showed hours to humans but told machines nothing.
How do I check in under a minute?
Homepage → right-click → View page source → Ctrl+F → search openingHours. Nothing found means nothing machine-readable.
Will this help my clinic show up in ChatGPT or AI Overviews?
Structured data doesn't guarantee inclusion in AI answers, but untruths and guesses are what happen in its absence. Facts in machine-readable form are the strongest, cheapest signal you control.
Shouldn't a NZ agency do this — why an overseas team?
Hours and schema don't care about borders; honesty does. We say where we are (India), work NZ-friendly hours, publish our method daily, and your first fix is free so you can judge the work before you ever judge the geography.
What exactly does the free first fix include?
One schema block for your clinic — name, address, phone, hours, service list — written for your site, plus a before/after verification you can run yourself in Google's own tester. No call required, no obligation, no follow-up chase.