The receptionist I built before anyone asked me to
An Acuity Edge case study. Anonymised for privacy: a 4.9-star, three-location independent veterinary group in New England, family-owned for over forty years.
Most automation pitches start with a slide deck. This one started with a call I made as a paying client, on hold, for over half an hour.
I'd taken my own animals to this practice for two years before I thought of it as a prospect. Three locations, 4.9 stars across more than 1,700 reviews, AAHA accredited, independent since the eighties. Clinically, one of the best in the region. The front desk was the only part not keeping pace, and I felt it as a client long before I diagnosed it. So I diagnosed it properly, built the fix, then went to them. Not the other way round.
What I found
I treated my own practice like a paid engagement: called every location, busy days and quiet, in hours and after; read the reviews, the job listings, the competitor sites. Six pain points, each verified myself rather than inferred. A genuine emergency sitting in the same queue as a routine vaccine reminder. Hold times past thirty minutes. No after-hours capability at all. Three locations running as three separate experiences, so a client who needs one site gets booked into another. A call-centre role advertised at salary rates to solve what is really a technology problem. And a competitor under a mile away already offering real-time online booking.
None of it is a knock on the people. A 4.9-star practice short-staffed on phones is one growing faster than its front desk can absorb. A good problem wearing a bad disguise.
What I built
Before I sent a word of outreach, I built Kitty, our AI receptionist, and trained her on this practice specifically. Not a generic demo. She knew the locations, the hours, the urgent-care routing, the refill process, and how to carry an end-of-life conversation with the warmth it demands. Building first removes the thing every owner braces for: a salesperson describing software they've never seen work on their own business. I didn't want them to imagine it. I wanted them to pick up the phone and talk to it.
The honest numbers
I built the case on the practice's own public figures and kept it deliberately conservative: one call-centre hire eliminated, a thirty per cent cut in missed-call losses rather than a hundred, one avoided turnover. No after-hours upside, no multi-location multiplier. Roughly $180,000 in first-year savings, payback under a month. I could have made the headline far bigger by counting every missed call as a client lost for life. I didn't, because a number you can't defend in the room is worth less than a smaller one you can. Structured as a thirty-day pilot at one location: clean exit, no obligation.
The lesson
The technology was never the hard part. Standing Kitty up took days. The hard part is that putting an AI in front of clients already on edge about a sick animal is a change-management problem, not a software one. It touches the receptionist who fears for her job, the owner who fears for the practice's voice, and the client who fears talking to a machine in a frightening moment. That's the work I actually do. The AI is just the delivery mechanism; making adoption feel safe is psychology before it's tech.
As it stands, the proposal is with them and I'm waiting on the final decision-maker to green-light the pilot. A forty-year practice doesn't rebuild its front desk overnight, and the people who sign that off are right not to rush it. What I'll stand behind either way is the method: diagnose it as a client would feel it, build the fix before I ask for anything, price it on numbers you can check yourself. If your phones are the weakest part of an otherwise strong practice, that's a twenty-minute conversation worth having.
Alexandra Whitfield, Acuity Edge