The booking request that landed in a screen nobody opened: a notify pattern for service businesses

A lot of Glide apps for local services (HVAC, clinics, cleaning) have a clean “Request a visit” form. The row appears in the Requests table, and then nothing happens. The owner is on a job, the office phone is ringing, and the new row sits unread until evening.

Here’s the pattern I use, with native Glide actions where possible:

  1. The form writes one row: Name, Phone (digits only), Issue, Preferred window, Status = New, Created = now.
  2. On submit, an action notifies the on-duty person right away by email or push, whichever that shop already checks. Keep the message short: “New request: Jane, AC not cooling, wants today. Open Requests.”
  3. Status is a single choice the owner can change from their phone: New, Called, Booked, No answer or Spam.
  4. A second, delayed path: if Status is still New after 15 minutes during business hours, notify a backup person. Outside hours, skip the ping and put the row at the top of a “Call first thing” list.
  5. Deduplicate: before creating a row, look up the phone number in the last 30 minutes. If it’s there, bump a “Calls” counter on that row and notify once more (“called again, urgent”) instead of creating three identical requests.

What makes it stick is that changing Status has to be one tap on the phone. If the owner needs a laptop, the escalation fires for jobs that were already handled.

Question for builders here: for the 15-minute “still New” check, do you use a scheduled workflow, or is there something simpler in Glide I’m missing?

You can use a scheduled workflow that runs every 5 to 10 minutes if your plan permits that cadence. It finds requests where:

  • Status = New
  • Created <= Now - 15 minutes
  • EscalationSent != true
  • Current time is within business hours

Correct. It will cost you minimum 1 update per interval (1 just to check if there are new calls…).

Thanks Mazen, that filter set is exactly what I was missing, especially the EscalationSent flag so the backup person doesn’t get pinged on every run. And good point on cost, Robert. Since each run burns at least one update even when nothing is new, I’m going to run it every 10 minutes and only inside business hours, so the schedule isn’t eating updates overnight. Has either of you tried having the submit action stamp an “escalate at” time on the row, so the workflow only touches rows where that time has passed? Curious whether that saves anything, or if the check itself is what counts.

Happy to help, And welcome to Glide Community :rose:

I’d stamp an EscalateAt time on submission. Makes filtering much cleaner. But I don’t think it saves Updates by itself; that depends on which workflow actions actually consume them. The real saving is avoiding unnecessary processing.

Thanks Mazen, and thanks for the welcome! That makes sense. I’ll stamp EscalateAt on submit for the cleaner filter and stop expecting it to save Updates on its own. I’m going to watch usage on a few runs that match zero rows, so I can see whether the scheduled check itself is what counts. I’ll post back here if I find anything useful.

It does. All server side workflows use a minimum of one update every time they are executed. You can calculate how many updates are used by any single execution by observing the graphic in the runflow log. Any step with a small blue lightning bolt next to it will use an update each time it is called. In the example below, a total of 2 updates are used:

Thanks Darren, that settles it. So even a run that matches zero rows costs one update, and every step with the lightning bolt adds another on top. At a 10-minute cadence that’s around 144 updates a day just for the check, so I’ll test whether 15 or 20 minutes is still fast enough for the escalation to be useful. Really helpful screenshot, I hadn’t noticed those bolts in the runflow log.