Freed's Bakery in Las Vegas makes custom cakes for weddings, birthdays and anniversaries, so they felt they had to answer calls at all hours. A phone tree didn't sort simple questions from real problems. Now an AI trained on each location's hours finds out what the caller needs and passes it on with context.
Business type: Restaurants & cafés. Tool used: Upfirst (AI receptionist). Result: The owner no longer fields calls in the middle of the night, and every after-hours call now comes with its reason.
Source: Upfirst case study (Company case study) · Added Tue, Oct 20 · Last updated 2026-10-20
The owner no longer fields calls in the middle of the night, and every after-hours call now comes with its reason. (as reported by the source)
Guide: AI receptionist for small business: real examples and how to set one up
Our own prompt for copying this setup. Swap in your details. Not quoted from the source.
You answer after-hours calls for [Bakery]. Answer hours and location questions from our info. If someone has a problem with a cake for an event in the next 48 hours, take their details and mark it URGENT. For everything else, take a message and say we'll call back when we open.
A voice agent takes phone orders from a live menu, suggests add-ons, adds tax and reads back the total, then the order appears in a Google Sheet the kitchen watches.
A Dallas fast-casual Mexican restaurant gets 45% of its sales by phone, and about 85% of callers order in Spanish. Staff had to choose between the line and the phone, losing 10 to 15% of phone orders. With a Spanish and English voice agent, the owner says phone orders tripled.
Rayo's Pizza in Hayward, California lost about 3 to 4 orders a day when a second call came in and callers hung up on hold. An AI now answers up to 10 calls at once and takes payment over the phone, so pickup is grab and go. It took 60 Super Bowl orders in 5 hours.