Our standard is: during the day, live answer or a callback inside 15 minutes, and the tech on the truck is never the one expected to pick up. That alone fixed half of it, because a guy under a sink answering with wet hands is worse than a fast callback.
What made the biggest difference for us:
The nights and weekends were our leak. We ended up putting an AI receptionist on overflow and after hours so callers get answered, the job details get captured, and we review in the morning. Here's the one we use if anyone's comparing: https://firstdeploy.ai
DanG, giving the callback list an owner and keeping the tech off the phone makes a lot of sense. The instant text also gives the customer a way to keep things moving while the office is busy.
One thing I’d add to the after-hours setup: test what happens when the on-call person doesn’t answer. Capturing an urgent request is only the first step; there needs to be a backup, and the caller shouldn’t hear “someone is on the way” unless that’s actually confirmed.
For anyone setting up an AI receptionist, I’d test three calls before going live: a routine estimate, an urgent service issue, and someone asking for a human. Check where each lands, who gets notified, and what the caller is promised.
For context, I’m the founder of Lacy.ai. Those are the checks I’d recommend with any provider.
Does your after-hours setup transfer urgent calls live, or alert someone to call back?
