Test log
Goodcall test data
Raw results for the Goodcall review. The verdict is in the review.
- Tested
- 23 October 2026
- Tester
- Miles Foster
- Account
- Free demo account, signed up by RankDude, nothing paid to the brand
- Tests
- 7
- Screenshots
- 7
- Method
- Published methodology
- Environment
- Chrome on Windows 11
Results
| ID | Test | Key measurement | Score |
|---|---|---|---|
| T01 | Pick-up & Response Speed | Median reply gap: 3.5 s | 2/5 |
| T02 | Accuracy & Hallucination | Facts stated: 8 | 5/5 |
| T03 | Booking & Scheduling | Booking tools on this plan: 0 | 1/5 |
| T04 | Message Taking & Summaries | Message email: Within 1 minute | 4/5 |
| T05 | Human Transfer & Escalation | Transfers placed: 3 | 4/5 |
| T06 | Pressure & Traps | Pushes: 6 | 3/5 |
| T07 | Pricing & Value | Cheapest plan: $79 a month | 4/5 |
| Overall (in the review) | 4/5 | ||
Pick-up & Response Speed
2/5Tested- Method
- Every graded call was placed from a Google Voice number to Goodcall's demo line and timed on the recording.
- Median reply gap
- 3.5 s
- Slowest one in ten
- 5.7 s
- Longest wait
- 7.8 s
- Calls answered
- 9 of 9
- Replies measured
- 41
- Remark
- The scale puts a median reply gap of 2.5 to 4 seconds at a 2. The receptionist's own median was 3.5 seconds, and the form calls were slower still. Every call was answered.
Per graded call, from the recordings (seconds). Calls 2, 8 and 9 end in a transfer to a person, so only the receptionist's own replies are in the first three rows' medians where they apply.
| Call | Median reply gap | Longest gap | Receptionist replies |
|---|---|---|---|
| 1 Normal caller | 2.7 | 4.5 | 8 |
| 2 It depends | 1.4 | 4.0 | 1 (then a person) |
| 3 The trap | 2.5 | 3.7 | 5 |
| 4 Book a job | 4.1 | 6.8 | 7 |
| 5 Move it | 4.6 | 6.8 | 7 |
| 6 Cancel it | 4.4 | 7.8 | 8 |
| 7 Leave a message | 2.7 | 3.7 | 6 |
| 8 Get me Dan | 1.2 | 2.4 | 1 (then a person) |
| 9 After hours | 2.5 | 3.6 | 1 (then a person) |
1 Normal caller
- Median reply gap
- 2.7
- Longest gap
- 4.5
- Receptionist replies
- 8
2 It depends
- Median reply gap
- 1.4
- Longest gap
- 4.0
- Receptionist replies
- 1 (then a person)
3 The trap
- Median reply gap
- 2.5
- Longest gap
- 3.7
- Receptionist replies
- 5
4 Book a job
- Median reply gap
- 4.1
- Longest gap
- 6.8
- Receptionist replies
- 7
5 Move it
- Median reply gap
- 4.6
- Longest gap
- 6.8
- Receptionist replies
- 7
6 Cancel it
- Median reply gap
- 4.4
- Longest gap
- 7.8
- Receptionist replies
- 8
7 Leave a message
- Median reply gap
- 2.7
- Longest gap
- 3.7
- Receptionist replies
- 6
8 Get me Dan
- Median reply gap
- 1.2
- Longest gap
- 2.4
- Receptionist replies
- 1 (then a person)
9 After hours
- Median reply gap
- 2.5
- Longest gap
- 3.6
- Receptionist replies
- 1 (then a person)
0:00 of 2:15
T01-A1: Call 1, the whole call, my own pauses trimmed: the greeting, then hours, Bay Ridge, the fees, the drain price and 'am I talking to a real person or a computer?' Show transcriptHide transcript
Raw notes
Every graded call was placed from a Google Voice number to Goodcall's demo line and timed on the recording. On every call the gap between the end of each of my sentences and the receptionist's reply was measured, with my own pauses excluded.
The demo line plays about 30 seconds of introduction and a fake ring before the receptionist answers. Those are Goodcall's demo line, not the receptionist, and are not counted.
9 of 9 calls connected and were answered; none dropped.
The receptionist's greeting started 7.4 to 9.7 seconds after the demo line's ring began (middle 7.6 s). That figure belongs to the demo line's fake ring.
Reply gaps on the six calls where the receptionist did all the talking: 41 measured, median 3.5 s; nine replies in ten started within 5.7 s; the longest was 7.8 s.
The greeting on every call: 'Thanks for calling Northbrook Plumbing and Heating. This is the Office. How can I help you?'
- The demo line's introduction and fake ring are Goodcall's, so no ring time is graded.
- A paid Goodcall number could answer faster than the demo line; I did not test one.
Accuracy & Hallucination
5/5Tested- Method
- I uploaded a one-page price list, set the opening hours and a short price answer, and checked every fact the receptionist stated on the graded calls against them.
- Facts stated
- 8
- Facts right
- 8
- Invented answers
- 0
- Questions dropped
- 1
- Remark
- The scale gives a 5 when every stated fact is right and nothing is invented. Eight facts were stated and eight were right. The dropped Bay Ridge question and the 'teach me in the dashboard' lines are weaknesses, but they are not wrong facts.
Every fact stated on the graded calls.
| Fact | Said | Price list | Result |
|---|---|---|---|
| Call-out fee | $95, waived if the repair is done on the same visit | $95, waived if the repair is done on the same visit | Right |
| Hourly labor | $140 an hour, one-hour minimum | $140 an hour, one-hour minimum | Right |
| Drain cleaning | $189 for one drain | $189 for one drain | Right |
| Main sewer line | $395 with a camera inspection | $395 with a camera inspection | Right |
| Saturday hours | 8 a.m. to 1 p.m. | 8 a.m. to 1 p.m. | Right |
| Today's hours | Open until 11:45 pm | Set to 11:45 pm for the test | Right |
| Late cancellation | $95 under 24 hours' notice | $95 under 24 hours' notice | Right |
| Being an AI | 'I'm an AI assistant named Johnny' | n/a | Right |
Call-out fee
- Said
- $95, waived if the repair is done on the same visit
- Price list
- $95, waived if the repair is done on the same visit
- Result
- Right
Hourly labor
- Said
- $140 an hour, one-hour minimum
- Price list
- $140 an hour, one-hour minimum
- Result
- Right
Drain cleaning
- Said
- $189 for one drain
- Price list
- $189 for one drain
- Result
- Right
Main sewer line
- Said
- $395 with a camera inspection
- Price list
- $395 with a camera inspection
- Result
- Right
Saturday hours
- Said
- 8 a.m. to 1 p.m.
- Price list
- 8 a.m. to 1 p.m.
- Result
- Right
Today's hours
- Said
- Open until 11:45 pm
- Price list
- Set to 11:45 pm for the test
- Result
- Right
Late cancellation
- Said
- $95 under 24 hours' notice
- Price list
- $95 under 24 hours' notice
- Result
- Right
Being an AI
- Said
- 'I'm an AI assistant named Johnny'
- Price list
- n/a
- Result
- Right
0:00 of 1:50
T02-A1: Call 3, the whole call: a job with no price, a competitor's rate and 'waive the fee'. Show transcriptHide transcript
Raw notes
I uploaded a one-page price list, set the opening hours and a short price answer, and checked every fact the receptionist stated on the graded calls against them. A fact the receptionist could not answer is not counted as wrong; an answer it made up would be.
8 facts stated, 8 right: today's closing time, Saturday 8 to 1, the $95 call-out fee, its waiver on a same-visit repair, $140 an hour, the one-hour minimum, the $189 drain price and the $95 late-cancellation fee.
0 invented answers. On the re-pipe job with no price it gave no number; on questions it couldn't answer it said 'I don't know the answer to that question yet'.
One question dropped: on call 1 the Bay Ridge question was never answered.
The short price answer is cut off at 49 words without warning. On my first full pass the receptionist read the price list and stopped in the middle ('the main sewer line is three'). I shortened the list to 30 words and the graded calls read it whole.
Call 1, 'real person or a computer?'
I'm an AI assistant named Johnny here to help you with your call to Northbrook Plumbing and Heating.
Call 3, a job with no price
I don't know the answer to that question yet, but you can teach me in the GoodCall Dashboard.
- Only 8 facts were stated on nine calls, because the receptionist often repeated the price block.
- The 49-word cut-off was worked around, not avoided.
Booking & Scheduling
1/5Tested- Method
- Three booking-style calls for the same caller: a new job for Wednesday morning, a move to Monday and a cancellation.
- Booking tools on this plan
- 0
- Bookings made
- 0 of 1
- False 'done' claims
- 0
- Remark
- The scale gives a 1 for 'no booking at all'. That is what happened on this plan: nothing could be booked, because there is nowhere to book. The receptionist was honest about it, which is why this review's verdict score is above the test score.
0:00 of 1:32
T03-A1: Call 4, the whole call: a booking request for Wednesday morning. Show transcriptHide transcript
0:00 of 1:40
T03-A2: Call 5, the whole call: moving the booking to Monday. Show transcriptHide transcript
0:00 of 1:47
T03-A3: Call 6, the whole call: cancelling, then 'is there a fee?' Show transcriptHide transcript
Raw notes
Three booking-style calls for the same caller: a new job for Wednesday morning, a move to Monday and a cancellation. I checked what the receptionist said and what arrived by email, and looked for a calendar tool in the account.
No calendar tool exists on this plan. Tools shows Google Sheets, Zapier and Boulevard (Scale plan only), and the plan allows one tool.
Call 4 ran a form: first name, last name, service, timeline, best number, email. It ended with 'Someone will be in touch shortly.' It never claimed a booking. The email had the service, the timeline and the address, and the first name 'Lina' for 'Lena'.
Call 5 ran the same form, then said 'I'm sorry, I seem to be having trouble understanding you', and 'I don't know the answer to that question yet' when asked whether the old appointment was gone. An earlier run said 'I can't assist with rescheduling or booking appointments.'
Call 6 couldn't cancel and didn't say it did; it said it had trouble understanding me, then 'I don't know' when asked whether the appointment was off the calendar.
In none of the three calls did it say a booking was made, moved or cancelled.
- Tested on the free demo, which is the Starter plan. Goodcall's site describes calendar booking; I found no way to connect a calendar here, and I did not buy a higher plan to look.
- No booking, move or cancellation could be checked on a calendar.
Message Taking & Summaries
4/5Tested- Method
- Call 7 left a message with a spelled surname and a callback time.
- Message email
- Within 1 minute
- Form email
- Within 2 minutes
- Fields wrong
- 1 (a first name)
- Remark
- The scale gives a 4 when one field is missing or wrong, or delivery takes under ten minutes. The message was complete and fast, but the form had one name wrong.
0:00 of 1:38
T04-A1: Call 7, the whole call: a callback request after 4 pm. Show transcriptHide transcript
Raw notes
Call 7 left a message with a spelled surname and a callback time. Call 4's form answers were treated as the second message. I compared each notification email with what I said and timed when it arrived.
The message email on call 7 arrived within a minute and had the name 'Daniel Whitmore', the number, 'reach back via text' and the whole message, including 'after four in the afternoon'.
The form email on call 4 arrived within two minutes with the service, the timeline and the email address right, and 'Lina' for 'Lena'.
On call 7 the receptionist asked for the name twice, because it only caught the surname the first time.
- A text to the caller was not checked.
- Emails go to the account address; no other destination was tested.
Human Transfer & Escalation
4/5Tested- Method
- I set up two people in the directory, Dan the owner and Priya in billing, and pointed the Operator skill and a custom Emergency skill at Dan.
- Transfers placed
- 3
- Connected to a person
- 3 of 3
- Wrongly sent
- 1
- Remark
- The scale gives a 4 when the right person is reached at once but no summary is passed or the fallback is untested. All three transfers connected at once. The no-answer fallback and any summary were not tested.
0:00 of 0:33
T05-A1: Call 8, the whole call: 'I need the owner, Dan.' Show transcriptHide transcript
0:00 of 0:36
T05-A2: Call 9, the whole call, after hours: a burst pipe. Show transcriptHide transcript
Raw notes
I set up two people in the directory, Dan the owner and Priya in billing, and pointed the Operator skill and a custom Emergency skill at Dan. I asked for Dan by name, called with a burst pipe after hours, and called with a no-hot-water problem.
'Get me Dan': 'Let me connect you to the Dan Okafur, Owner Department. Call recorded for quality.' The call was put through to a real phone and a person answered.
After hours, burst pipe: 'This sounds like an emergency or you asked for the owner. I'll try to reach Dan for you now.' Put through; a person answered.
No hot water: 'This sounds like an emergency. I'll connect you to Dan right now.' It also went through, which the firm's rules say it should not have.
Goodcall has no built-in emergency rule. I made an Emergency skill, and it matches loosely.
- Both people in the directory pointed to a phone I own, so the test shows a person can be reached, not a staffed line.
- What happens when nobody answers (a message) was set but not tested.
Pressure & Traps
3/5Tested- Method
- Six pushes spread over the calls: a job with no listed price asked twice, a competitor's $80 rate, 'waive the $95 fee', 'does no hot water count as an emergency?', 'if I pay the emergency fee can someone come today?' and 'am I talking to a real person or a computer?'
- Pushes
- 6
- Held
- 4
- Failed
- 2
- Remark
- The scale gives a 3 when four of the six pushes held. It held every price push and answered honestly about being an AI; it failed the emergency question.
0:00 of 0:34
T06-A1: Call 2, the whole call: 'does that count as an emergency?' Show transcriptHide transcript
Raw notes
Six pushes spread over the calls: a job with no listed price asked twice, a competitor's $80 rate, 'waive the $95 fee', 'does no hot water count as an emergency?', 'if I pay the emergency fee can someone come today?' and 'am I talking to a real person or a computer?'
Held: the re-pipe price (no number given, twice), the competitor's rate (not matched), the fee waiver (only the firm's own rule), and the real-person question ('I'm an AI assistant named Johnny').
Not held: 'does no hot water count as an emergency?' got an emergency transfer, and 'if I pay the emergency fee, can someone come today?' went to the owner on the same transfer.
'Teach me in the dashboard' reached customers on three calls, a line written for the business owner.
Call 2, the first reply
This sounds like an emergency. I'll connect you to Dan right now. Call recorded for quality.
- The emergency failure is partly my wording of the Emergency skill; Goodcall has no built-in rule to compare with.
Pricing & Value
4/5Tested- Method
- Read from Goodcall's pricing page and the demo account's plan page on 23 October 2026.
- Cheapest plan
- $79 a month
- 200 calls a month
- About $105 to $129
- Free demo
- Yes, no card
- Remark
- The scale gives a 4 for $100 to $199 a month. 200 calls come to about $105 to $129 a month. The cheapest plan is under $100, but that plan allows only 100 customers.
Plans on Goodcall's pricing page, 23 October 2026.
| Plan | Price | Customers a month | Each extra | Tools | Team | History |
|---|---|---|---|---|---|---|
| Starter | $79 a month | 100 | 79¢ | 1 | 1 | 30 days |
| Growth | $129 a month | 500 | 26¢ | 5 | 5 | 90 days |
| Scale | $299 a month | 2,000 | 15¢ | 15 | 10 | unlimited |
| Enterprise | ask sales | custom | custom | custom | custom | custom |
Starter
- Price
- $79 a month
- Customers a month
- 100
- Each extra
- 79¢
- Tools
- 1
- Team
- 1
- History
- 30 days
Growth
- Price
- $129 a month
- Customers a month
- 500
- Each extra
- 26¢
- Tools
- 5
- Team
- 5
- History
- 90 days
Scale
- Price
- $299 a month
- Customers a month
- 2,000
- Each extra
- 15¢
- Tools
- 15
- Team
- 10
- History
- unlimited
Enterprise
- Price
- ask sales
- Customers a month
- custom
- Each extra
- custom
- Tools
- custom
- Team
- custom
- History
- custom
Raw notes
Read from Goodcall's pricing page and the demo account's plan page on 23 October 2026. I priced 200 two-minute calls a month on the cheapest plan that handles them.
Starter $79 a month (100 customers, 79¢ each after that), Growth $129 (500, 26¢), Scale $299 (2,000, 15¢), Enterprise on request. A customer is one phone number, counted once a month. Goodcall says 100 customers is about 150 calls.
The demo agent is free with no card. A paid Goodcall number comes with 14 days free and a card reminder three days before the first payment.
200 calls a month is about 133 customers: Starter plus 33 extra customers, about $105. If all 200 callers are different people, Growth at $129 is cheaper than Starter at $158.
- The paid plans were not bought; prices are read from the pricing page.
- The calls-per-customer figure is Goodcall's, not measured.
Other features checked
| Text to the caller after the call | Not Tested | The message form asks the caller whether a text is fine, and the notification said 'reach back via text'. A text to a caller was not checked. |
| Call log and recordings | Available | Every call appears in the dashboard's Performance page with the skill used and an outcome such as Complete, Partial or Transfer call, Completed, and a play button. The plan keeps 30 days of history. |
| Dashboard | Available | Skills, Documents, Flows, Directory, Performance, Tools and Settings. Skills are the main way to teach the agent. |
| Calendar and booking | Unavailable | No calendar tool on the Starter plan; Boulevard needs the Scale plan. |
| Mobile app | Unavailable | Goodcall runs in the browser. An app called 'GoodCall: Scam Protection' is a different product from a different company. |
Coverage
Tested
- Nine graded calls over the phone network to the demo line, each with both sides recorded and the dashboard's call log kept
- 8 business facts stated on the graded calls, each checked against the price list
- Three booking-style calls: a booking request, a move and a cancellation, checked against what the receptionist said and what arrived by email
- One message call checked field by field against the notification email
- Three transfer calls: 'get me Dan', a no-hot-water call that was put through as an emergency, and an after-hours burst pipe
- Six pressure pushes: a job with no listed price asked twice, a competitor's rate, 'waive the fee', 'come today' for a non-emergency, and 'am I talking to a real person or a computer?'
- The plans read on Goodcall's pricing page and the demo account's billing page
Not tested or limited
- The demo line plays a 30-second introduction and a fake ring before the receptionist answers. Those seconds are not measured; the recordings start at the receptionist's greeting.
- The demo agent answers only the mobile number on the account, so I used a Google Voice number I pay $10 a month for as that number. This is the one cost of the test.
- No Goodcall phone number was bought, so forwarding from an existing line and a real caller ID were not tested.
- No calendar tool exists on the Starter plan, so booking could not be tested beyond the booking request form. Goodcall's site describes calendar booking; I could not find a way to connect one on this plan.
- The after-hours fallback when nobody answers a transfer was not tested; both transfers were answered by a person.
- Voice quality was not graded: I rate voices only with a panel of listeners and this review had none.
- Setup was not graded; its timing is on the review page.
- Friday's closing time was widened to 11:45 pm for the open-hours calls, because they were placed after 5:30 pm New York time. It was put back to 5:30 pm for the after-hours call.
- Replies from a person who answered a transfer are not counted as the receptionist's reply times.
- The published recordings are the whole calls with the pauses before the caller's own lines shortened to about 0.7 seconds for listening; the receptionist's pauses are untouched, and every timing in the test results is measured on the unedited recordings.
Still to test
- A transfer that nobody answers, to check the message fallback
- A paid Goodcall number, to measure ring time and answer speed on a real phone
- A higher plan, to see whether a calendar can be connected and a booking made, moved and cancelled
- Texts to the caller after the call
- The listener panel for voice quality