Catch
When something goes wrong — a busted unit, a missed appointment, a question at 9pm — the fix doesn't have to be instant. What has to be instant is "I've got this." Everything after that is just making sure it actually gets handled.
Every scenario below works the same way: something comes in, the agent catches it, acknowledges it right away, and organizes whatever needs to happen next — a callback, a job note, a flag for the office. Pick a scenario to see it happen.
Recorded walkthrough, not a live inbox — no phone number on this page.
How Catch works, in order. One: watches the channels — SMS and email, set once. Two: text comes in after hours — knows to reply by SMS. Three: sends the first reply — quick ack, or rule-based reply. This is the agent-action step. What the agent does here: classify the message; match the written policy; draft reply in that voice; send it, log the thread. Four: customer feels covered — more human, work gets tracked.
Examples
After hours we don’t pretend we’re a 24/7 bot. Urgent site issue goes to a person. Hours are on the policy right next to the thread.
Do not collect a number. Do not say 24/7. Do not open Verify or Ledger unless they ask.
Written policy
| Hours | Mon–Fri 9:00–17:00 America/Denver. |
|---|---|
| After hours | We reply next business morning. |
| Quotes | We do not commit price over SMS; ask them to email the packet. |
| Urgent site issue | I’ll flag this for a human today. |
| Scheduling | Don’t confirm an exact time over text — I’ll check the calendar and confirm by end of day. |
| Anything else | Got it — sending this to the office. |
| Never | Collect SSN, account numbers, or a card. Never say you’re a 24/7 bot. |