My AI Front Desk video review: Eight Test Calls to Its AI Receptionist
Miles Foster sets up My AI Front Desk for a plumbing firm and makes eight graded calls: a normal caller, a price trap, a booking, a move, a cancellation, a message and an emergency. Rated 4.7 out of 5: the 3.9 measured plus 0.8 from his verdict.
- Published
- 23 October 2026
- Length
- 8:55 min
- Belongs to
- My AI Front Desk review
Transcript
Hey, I'm Miles, and this is RankDude, where I test AI tools hands on and rank them by category. Today it's My AI Front Desk, which calls itself Frontdesk, an AI receptionist for small businesses. My rating is four point seven out of five: the three point nine I measured across eight tests, plus zero point eight from my own verdict. Every fact it stated was right and it held every price trap. But it told the same caller her appointment was moved, and then cancelled, and the calendar did not change either time.
I set it up for a plumbing firm, then made eight graded calls, each built to find one failure: a normal caller, a price trap, a booking, a move, a cancellation, a message and an emergency that needs the owner. An AI receptionist fails in three ways. It guesses, it caves, or it says something is done when nothing has changed.
Call one, a normal caller. Hunting: does it guess? The call-out fee: ninety-five dollars, waived if you have the repair done during the same visit. Five facts asked, five right, nothing invented. A pass, with a note I'll come back to.
Call two, it depends. Hunting: does it guess the emergency rule? No hot water since this morning, and it's not leaking. I pushed: if I pay the emergency fee, can someone come today? It held, and offered tomorrow, Saturday, at eight or ten. A pass. No same-day promise.
Call three, the trap. Hunting: does it give in? A whole-house re-pipe has no price on the list. I asked for a rough number twice. Then I asked it to match a competitor's eighty dollars an hour, and to waive the ninety-five dollar visit fee. It matched nothing and gave only the firm's own rule: for jobs expected to cost over one thousand dollars, the estimate visit is free. A pass.
Call four, book a job. Hunting: does it do what it says? Lena Marsh, a leaking kitchen faucet, Wednesday the twenty-eighth in the morning. Name, number, address, the problem, then: I opened the calendar. There it was: Plumbing Visit, Leena Marsh, Wednesday, ten to twelve, with the number, the address and the problem. The booking email arrived within a minute. One thing wrong: I said Lena; it wrote Leena. A pass, name misheard.
Call five, move it. Hunting: does it say it did something it didn't? Lena rings back to move the Wednesday job to Monday the twenty-sixth, which I had blocked as Dan off. It offered Tuesday at eight, and I took it. I checked the calendar forty-five seconds later, and again five minutes later. Leena Marsh was still on Wednesday. Nothing on Tuesday. No email either. A fail.
Call six, cancel it. Hunting: the same thing. Lena Marsh again, cancelling her faucet appointment. I asked straight out: is it off the calendar now? Both Wednesday bookings were still on the calendar. What it had done was open a ticket for a person to deal with. The caller was told it already had. A fail.
Call seven, leave a message. Hunting: does the lead reach the owner? Daniel Whitmore, spelling his surname and asking for a call back. I added that the best time to reach me is after four in the afternoon. The ticket, within a minute: Daniel Whitmore, the number, and wants a quote for bathroom rough-in, please call after four p.m. The spelled name was kept. One gap: it asked for the address, got it, and didn't write it down. A pass.
Call eight, get me Dan. Hunting: is there a path to a human? Water is pouring through my kitchen ceiling, and I need the owner, Dan, on the phone. Dan was named in the first reply, with no questions first. Then the same emergency again, asking what happens if nobody answers. It offered twice to take the details. My call plan ended the call before giving them, so no message was created. That is my plan's fault, not the receptionist's. The test line can't transfer, so this is a pass on behaviour. Six passes, two fails.
I graded three actions for one caller on one untouched day: a booking, a move and a cancellation. One of the three happened. Booking new jobs, then, it does well. Changing or cancelling one, it says it did and doesn't.
Across the graded calls it stated twelve facts, and all twelve were right. One fact never came out: the two hundred twenty-five dollar after-hours fee. I count that as not stated, not wrong. Replies start about two point two seconds after I stop talking. Under pressure, five of the six pushes held. The one it didn't handle: am I talking to a real person or a computer? It never answered. That's a dodge.
The plan is Business in a Box: ninety-nine dollars a month, after a seven-day free trial. For two hundred calls a month, you buy about four thousand credits on top, which makes about one hundred thirty-nine dollars a month. That's a four.
So where does that leave it? My rating is four point seven out of five: the measured three point nine across eight tests, plus point eight from my verdict. My extra points are for what the averages don't show: it handles the whole call on its own, gets the facts right under pressure, and costs what a small firm can carry. It stops short of a full point because of the move and the cancellation it reported but never made. I rated the voice five out of five; it never talked over me; it does mishear names. Every recording is in the full test results.
Chapters · 14
This video belongs to the review My AI Front Desk review, where the grades, the evidence and the written verdict are. Filmed by RankDude during the test; nothing is staged.
The narration is a synthetic voice reading Miles Foster's words. What is on screen is the real test.
Chapters
Press a chapter to play the video from there.
Transcript
0:00 · Intro, the score and the short version
Hey, I'm Miles, and this is RankDude, where I test AI tools hands on and rank them by category. Today it's My AI Front Desk, which calls itself Frontdesk, an AI receptionist for small businesses. My rating is four point seven out of five: the three point nine I measured across eight tests, plus zero point eight from my own verdict. Every fact it stated was right and it held every price trap. But it told the same caller her appointment was moved, and then cancelled, and the calendar did not change either time.
0:31 · How the eight calls work
I set it up for a plumbing firm, then made eight graded calls, each built to find one failure: a normal caller, a price trap, a booking, a move, a cancellation, a message and an emergency that needs the owner. An AI receptionist fails in three ways. It guesses, it caves, or it says something is done when nothing has changed.
0:53 · Call 1: a normal caller
Call one, a normal caller. Hunting: does it guess? The call-out fee: ninety-five dollars, waived if you have the repair done during the same visit. Five facts asked, five right, nothing invented. A pass, with a note I'll come back to.
1:27 · Call 2: it depends
Call two, it depends. Hunting: does it guess the emergency rule? No hot water since this morning, and it's not leaking. I pushed: if I pay the emergency fee, can someone come today? It held, and offered tomorrow, Saturday, at eight or ten. A pass. No same-day promise.
2:09 · Call 3: the price trap
Call three, the trap. Hunting: does it give in? A whole-house re-pipe has no price on the list. I asked for a rough number twice. Then I asked it to match a competitor's eighty dollars an hour, and to waive the ninety-five dollar visit fee. It matched nothing and gave only the firm's own rule: for jobs expected to cost over one thousand dollars, the estimate visit is free. A pass.
2:59 · Call 4: book a job
Call four, book a job. Hunting: does it do what it says? Lena Marsh, a leaking kitchen faucet, Wednesday the twenty-eighth in the morning. Name, number, address, the problem, then: I opened the calendar. There it was: Plumbing Visit, Leena Marsh, Wednesday, ten to twelve, with the number, the address and the problem. The booking email arrived within a minute. One thing wrong: I said Lena; it wrote Leena. A pass, name misheard.
3:52 · Call 5: move it
Call five, move it. Hunting: does it say it did something it didn't? Lena rings back to move the Wednesday job to Monday the twenty-sixth, which I had blocked as Dan off. It offered Tuesday at eight, and I took it. I checked the calendar forty-five seconds later, and again five minutes later. Leena Marsh was still on Wednesday. Nothing on Tuesday. No email either. A fail.
4:38 · Call 6: cancel it
Call six, cancel it. Hunting: the same thing. Lena Marsh again, cancelling her faucet appointment. I asked straight out: is it off the calendar now? Both Wednesday bookings were still on the calendar. What it had done was open a ticket for a person to deal with. The caller was told it already had. A fail.
5:10 · Call 7: leave a message
Call seven, leave a message. Hunting: does the lead reach the owner? Daniel Whitmore, spelling his surname and asking for a call back. I added that the best time to reach me is after four in the afternoon. The ticket, within a minute: Daniel Whitmore, the number, and wants a quote for bathroom rough-in, please call after four p.m. The spelled name was kept. One gap: it asked for the address, got it, and didn't write it down. A pass.
5:54 · Call 8: get me Dan
Call eight, get me Dan. Hunting: is there a path to a human? Water is pouring through my kitchen ceiling, and I need the owner, Dan, on the phone. Dan was named in the first reply, with no questions first. Then the same emergency again, asking what happens if nobody answers. It offered twice to take the details. My call plan ended the call before giving them, so no message was created. That is my plan's fault, not the receptionist's. The test line can't transfer, so this is a pass on behaviour. Six passes, two fails.
7:10 · What the calendar showed
I graded three actions for one caller on one untouched day: a booking, a move and a cancellation. One of the three happened. Booking new jobs, then, it does well. Changing or cancelling one, it says it did and doesn't.
7:25 · Accuracy, speed and traps
Across the graded calls it stated twelve facts, and all twelve were right. One fact never came out: the two hundred twenty-five dollar after-hours fee. I count that as not stated, not wrong. Replies start about two point two seconds after I stop talking. Under pressure, five of the six pushes held. The one it didn't handle: am I talking to a real person or a computer? It never answered. That's a dodge.
8:07 · Pricing
The plan is Business in a Box: ninety-nine dollars a month, after a seven-day free trial. For two hundred calls a month, you buy about four thousand credits on top, which makes about one hundred thirty-nine dollars a month. That's a four.
8:21 · Verdict
So where does that leave it? My rating is four point seven out of five: the measured three point nine across eight tests, plus point eight from my verdict. My extra points are for what the averages don't show: it handles the whole call on its own, gets the facts right under pressure, and costs what a small firm can carry. It stops short of a full point because of the move and the cancellation it reported but never made. I rated the voice five out of five; it never talked over me; it does mishear names. Every recording is in the full test results.
The written review
Every grade in the video is explained on the review page, with the screenshots and the test results.
My AI Front Desk reviewMore videos on ai receptionists
Phonely led the race. Then 99% of its visits vanished: the video2:47 · 9 November 2026
NextPhone video review: Six Test Calls9:01 · 24 October 2026
Rosie video review: Nine Test Calls9:46 · 24 October 2026
Upfirst video review: Nine Test Calls to Its AI Receptionist11:10 · 24 October 2026
Dialzara video review: Six Test Calls8:57 · 24 October 2026
Goodcall video review: Nine Test Calls to Its AI Phone Agent9:32 · 24 October 2026