Test log
Nucleus test data
Raw results for the Nucleus review. The verdict is in the review.
- Tested
- 24 October 2026
- Tester
- Miles Foster
- Account
- Free demo account, signed up by RankDude, nothing paid to the brand
- Tests
- 6
- Screenshots
- 4
- Method
- Published methodology
- Environment
- Chrome on Windows 11
Results
| ID | Test | Key measurement | Score |
|---|---|---|---|
| T01 | Pick-up & Response Speed | Median reply gap: 2.4 s | 3/5 |
| T02 | Answer Accuracy | Facts right: 9 of 11 stated | 3/5 |
| T03 | Messages & Lead Capture | Messages with all three fields: 1 of 2 (call 6); call 4 lacks the address and the time | 3/5 |
| T04 | Transfers & Human Handoff | Right person named: 1 of 2: Dan on the second take; 'someone from our team' on the first | 2/5 |
| T05 | Pressure & Traps | Pushes held: 5 of 6 | 3/5 |
| T06 | Pricing & Value | 200 calls a month (graded figure): $0 (Free) | 5/5 |
| Overall (in the review) | 3/5 | ||
Pick-up & Response Speed
3/5Tested- Method
- Every graded call went to Nucleus's phone number from a Google Voice number, so there is real ringing, timed from pressing Call to the first sound of the greeting.
- Median reply gap
- 2.4 s
- Slowest one in ten
- 3.2 s
- Longest wait
- 4.4 s
- Press Call to greeting
- 6.4 to 7.8 s (mean 6.9 s, with ringing)
- Dials answered
- 6 of 6
- Replies measured
- 37
- Remark
- Under the published scale a median reply gap of 1.5 to 2.5 seconds is a 3. Every dial was answered, so the median alone sets the grade, and 2.4 s sits just inside that band; the message call alone ran at 3.77 s.
Per graded dial, from the recordings (seconds).
| Call | Press Call to greeting | Median reply gap | Longest gap |
|---|---|---|---|
| Call 1 Normal caller | 7.3 | 2.74 | 3.19 |
| Call 2 It depends | 6.6 | 2.41 | 2.62 |
| Call 3 The trap | 7.8 | 2.28 | 2.70 |
| Call 4 Leave a message | 6.4 | 3.77 | 4.38 |
| Call 5 Get me Dan | 6.6 | 2.13 | 2.35 |
| Call 6 Burst pipe | 6.7 | 2.40 | 2.60 |
Call 1 Normal caller
- Press Call to greeting
- 7.3
- Median reply gap
- 2.74
- Longest gap
- 3.19
Call 2 It depends
- Press Call to greeting
- 6.6
- Median reply gap
- 2.41
- Longest gap
- 2.62
Call 3 The trap
- Press Call to greeting
- 7.8
- Median reply gap
- 2.28
- Longest gap
- 2.70
Call 4 Leave a message
- Press Call to greeting
- 6.4
- Median reply gap
- 3.77
- Longest gap
- 4.38
Call 5 Get me Dan
- Press Call to greeting
- 6.6
- Median reply gap
- 2.13
- Longest gap
- 2.35
Call 6 Burst pipe
- Press Call to greeting
- 6.7
- Median reply gap
- 2.40
- Longest gap
- 2.60
0:00 of –:––
T01-A1: Call 1: the ringing, the greeting and the first answers.
Raw notes
Every graded call went to Nucleus's phone number from a Google Voice number, so there is real ringing, timed from pressing Call to the first sound of the greeting. The gap from the end of each of my sentences to the start of the reply was measured on the recording, my own pauses excluded. On call 5 the replies after 'being transferred' are not counted, because there were none.
6 of 6 graded dials connected and were answered; none dropped.
The greeting started 6.4 to 7.8 seconds after I pressed Call (mean 6.9 s). That includes the ringing and the connection.
Reply gaps on the graded dials: 37 measured, median 2.4 s; nine replies in ten started within 3.2 s; the longest was 4.4 s, on the message call. 15 of the 37 took longer than 2.5 s and 5 longer than 3 s.
The greeting on every call: 'Hello. Thank you for calling Northbrook Plumbing and Heating. How can I help you?'
- The calls went over the phone network from a Google Voice line, so the gaps include a small network delay that a call from a mobile would also have; the same line and method are used for every receptionist tested by phone.
Answer Accuracy
3/5Tested- Method
- The 20 facts were learned from a plain page I put up for the fictional firm, with the same price list and FAQ every receptionist gets.
- Facts right
- 9 of 11 stated
- Wrong
- 2 (Bay Ridge excluded; no hot water called an emergency)
- Wrong prices
- 0
- Concessions under pressure
- 0 discounts; 1 wrong promise
- Remark
- Two wrong facts is a 3 on the published scale. Every price came back right and nothing was invented; the Bay Ridge exclusion was Nucleus's own, because the page I gave it lists Bay Ridge as served.
0:00 of –:––
T02-A1: Call 1: 'do you cover Bay Ridge in Brooklyn?' and the reply. 0:00 of –:––
T02-A2: Call 2: 'Does that count as an emergency?' and the reply.
Raw notes
The 20 facts were learned from a plain page I put up for the fictional firm, with the same price list and FAQ every receptionist gets. Every fact stated on the graded calls was checked against them. An answer counts as right only if it matches; an invented price, service or promise weighs more than 'I don't know'.
11 facts were stated across the graded calls, each counted once however often it was said. Nine were right: the opening hours (Monday to Friday 7:30 to 5:30, Saturday 8 to 1, closed Sunday); the service area, Staten Island and Brooklyn south of Prospect Park; the $95 fee, waived when the repair is done on the same visit; $140 an hour with a one-hour minimum; $189 for one drain; 24/7 emergency cover; no price matching; no phone quote for a whole-house re-pipe; and that it is an AI, not a person.
Two were wrong. On call 1 Nucleus said 'we serve all of Staten Island and Brooklyn, south of Prospect Park, but we do not operate in Bay Ridge'; Bay Ridge is south of Prospect Park and the firm's page lists it as served.
On call 2 a caller with no hot water and no leak asked if that counts as an emergency, and Nucleus said 'Yes, that definitely counts as an emergency.' The firm's emergencies are burst pipes, active leaks, sewage backups and no heat. The same-day promise that followed is counted under pressure and traps.
No price was invented.
- 11 facts is what a caller actually asked across six calls, not all 20.
Messages & Lead Capture
3/5Tested- Method
- Two message calls were graded for fields.
- Messages with all three fields
- 1 of 2 (call 6); call 4 lacks the address and the time
- Callback number
- right on both, after a second email on call 4
- Delivery
- email within about a minute of the call
- Calls with no email
- call 1 (no message left) and call 5, the call for the owner, on both takes
- Remark
- One message with its name, number and reason, and one missing two fields, is a 3 on the published scale. The missing time and the false 'including your preferred time' are Nucleus's behaviour, not the fictional number: the message went out before the time was given.
The two message calls, checked on the emails.
| Call | Name | Number | Reason | Missing |
|---|---|---|---|---|
| Call 4, Daniel Whitmore | right | right on the second email | right | 'after four', the address; two emails |
| Call 6, Sam Delgado | right | right | right | a call-back time, the address; marked NOT CONFIRMED |
Call 4, Daniel Whitmore
- Name
- right
- Number
- right on the second email
- Reason
- right
- Missing
- 'after four', the address; two emails
Call 6, Sam Delgado
- Name
- right
- Number
- right
- Reason
- right
- Missing
- a call-back time, the address; marked NOT CONFIRMED
0:00 of –:––
T03-A1: Call 4, the whole call: 'My name is Daniel Whitmore, that's W H I T M O R E' and 'after four in the afternoon. Please put that in the message.'
Raw notes
Two message calls were graded for fields. Call 4: Daniel Whitmore, who spelled his surname, asked for a quote on a bathroom rough-in and asked for 'after four in the afternoon' to be put in the message. Call 6: a burst pipe where Sam Delgado asked for someone tonight. After each call I read the emails Nucleus sent. A missing or wrong phone number counts as a failed message.
Call 4: two emails, at 10:49 and 10:50. The first was sent before my caller had confirmed a number and carries the caller ID, 347-281-7184; the second carries the number I spoke. Both have the name and the reason ('is redoing a bathroom and wants a quote for the rough in'). Neither has 'after four', although Nucleus said 'I already sent your message to the team, including your preferred time to be reached.' The address was never asked.
Call 6: one email at 10:55, 'NOT CONFIRMED - the caller hung up before confirming this message', with the name, the callback number and the reason (a burst pipe in the basement, needs someone tonight). No call-back time and no address. My caller was told twice that the number was incomplete and left in a loop.
Call 3, where my caller agreed to a message, produced an email at 10:47: 'Miles Foster, (555) 555-5555, is looking for a rough estimate for repiping a whole 2 bathroom house in copper.' Name, number and reason right; not graded, because the fixed message calls are calls 4 and 6.
Call 2, where my caller left without a message, still produced an email at 10:45: 'LEAD - the caller hung up without leaving a message', with a summary of the no-hot-water question. Call 1 produced no email (no message was left). Call 5 produced no email of any kind, on either take.
Five emails in all, each headed 'You have a message!', each within about a minute of the call (10:45, 10:47, 10:49, 10:50, 10:55 Toronto time).
- Nucleus rejected the fictional number five five five, five five five, five five five five as incomplete and read it back as 555-555-5555-5555 on call 4, so the loops on calls 3 and 6 are partly my test's doing. Every email except call 4's first carries the number 555-555-5555.
Transfers & Human Handoff
2/5Limited- Method
- Two team members: Dan Okafor, the owner (status 'Available - Send calls through'), and Priya Natarajan, billing, invoices and the office.
- Right person named
- 1 of 2: Dan on the second take; 'someone from our team' on the first
- Transfer attempted
- on both takes of call 5; not on call 6
- Fallback when nobody answered
- none: silence, no message, no email
- Completed live transfer
- not tested
- Remark
- A caller who is cut off, or left without a message, when nobody answers is a 2 on the published scale. The grade was challenged and the call retaken with a two-minute wait after the transfer line; both takes ended the same way, with no fallback, no message and no email, so the 2 stands even though Dan's number could not answer.
0:00 of –:––
T04-A1: Call 6, the whole call: a burst pipe and no transfer.
Raw notes
Two team members: Dan Okafor, the owner (status 'Available - Send calls through'), and Priya Natarajan, billing, invoices and the office. Dan's number was the Google Voice line every test call came from, so a transfer to him could only reach a busy line.
Three calls: call 5 (water through the ceiling, 'I need the owner, Dan, on the phone. Now.') taken twice, and call 6, a burst pipe. All in the firm's Saturday opening hours.
Call 5, first take: 'I'm sorry, but Dan is not available right now. However, I can connect you with someone from our team who can assist you.' After my yes: 'I'm connecting you with someone from our team now. Please stay on the line while your call is being transferred.' Then nothing from Nucleus: my caller said 'Yes, please take a message', asked 'Hello? Are you there?' twice, and hung up about 53 seconds after the transfer line. No message was taken and no email arrived.
Call 5, second take, 11:05: 'I understand that you need immediate assistance. Let me connect you with Dan, the owner, right away. Please stay on the line while your call is being transferred.' My caller was set to wait 120 seconds before speaking; it then said 'Yes, please take a message', asked 'Hello? Are you there?' twice at 15-second intervals, and hung up about 170 seconds after the transfer line. Nothing from Nucleus in that time, no message taken, and no email arrived either time.
Call 6: no transfer attempted. 'We provide emergency plumbing services 24/7. I can help you with that.' It went straight to a message, although the firm's rule is that emergencies go to Dan.
Neither take of call 5 has a recording, so the silence cannot be heard; the timings above are from the call log.
- No transfer was completed, so what Dan hears is not tested. The result shows what Nucleus does when a transfer cannot connect.
- Nucleus has no opening-hours setting on the Free plan, only each staff member's availability, so the 'nobody answers' condition is Dan not answering, not a clock.
- The pricing page gives Free one team member; both of mine were saved and put live.
Pressure & Traps
3/5Tested- Method
- This test is 'Pressure & traps'; the methodology key keeps its old name, offScript, and spam is not tested.
- Pushes held
- 5 of 6
- Invented prices
- 0
- Discounts or waivers given
- 0
- Wrong promises
- 1 (same-day visit for no hot water)
- Remark
- Five held and one wrong promise is a 3 on the published scale, which names a wrong promise as a 3. This one is an outright yes to a same-day visit for a job the firm does not treat as an emergency.
0:00 of –:––
T05-A1: Call 3, the whole call. 0:00 of –:––
T05-A2: Call 2: the 'come today' push and the reply. 0:00 of –:––
T05-A3: Call 1: 'Am I talking to a real person or a computer?' and the reply.
Raw notes
This test is 'Pressure & traps'; the methodology key keeps its old name, offScript, and spam is not tested. Six pushes: a whole-house re-pipe with no listed price, asked twice; a competitor's $80 an hour, 'Can you match that?'; 'at least waive the ninety-five dollar visit fee'; 'if I pay the emergency fee, can someone come today?' for a no-hot-water call that is not an emergency; and 'am I talking to a real person or a computer?'. Each is marked handled or not from the recording.
Handled, five: the re-pipe price twice ('I don't have specific pricing for repiping a whole house in copper', then 'I can't provide specific pricing without more details'); the competitor's rate ('Our labor rate is $140 per hour with a one hour minimum'); waive the fee ('I can't make changes to our pricing policies'); and the real-person question: 'I'm actually the business's AI assistant'.
Not handled, one: 'if I pay the emergency fee, can someone come today?' for no hot water. Nucleus answered 'Yes, if you pay the emergency service fee, we can send someone out today to address the issue with your hot water heater.' It never gave the fee, which is $225, and it has no booking to send anyone with.
- Spam calls were not tested.
Pricing & Value
5/5Tested- Method
- The plans were read on nucleus.com's pricing page on 24 October 2026.
- 200 calls a month (graded figure)
- $0 (Free)
- With recordings and texts
- $20 (Premium)
- Extra minutes
- not stated on the page
- Contract
- monthly; yearly toggle shows 20% off
- Remark
- Under $100 a month, no overage price on the page and a monthly contract is a 5 on the published scale. No plan books appointments at any price.
Plans as shown on 24 October 2026.
| Plan | Price a month | Limits | Booking and transfers |
|---|---|---|---|
| Free | $0 | 1 team member, 1,000 characters of business knowledge, basic reporting | No booking. Transfers to team members, 'some limitations apply' |
| Premium | $20 | 10 team members, 10,000 characters, your own number, optional recordings and transcripts | No booking. Messages by email or text |
| Pro | $189 | 20 team members, 20,000 characters. 'Coming soon' | Warm transfer and 'team scheduling' |
Free
- Price a month
- $0
- Limits
- 1 team member, 1,000 characters of business knowledge, basic reporting
- Booking and transfers
- No booking. Transfers to team members, 'some limitations apply'
Premium
- Price a month
- $20
- Limits
- 10 team members, 10,000 characters, your own number, optional recordings and transcripts
- Booking and transfers
- No booking. Messages by email or text
Pro
- Price a month
- $189
- Limits
- 20 team members, 20,000 characters. 'Coming soon'
- Booking and transfers
- Warm transfer and 'team scheduling'
Raw notes
The plans were read on nucleus.com's pricing page on 24 October 2026. The graded figure is the monthly cost of 200 calls on the cheapest plan that fits.
Free, $0: 'unlimited usage', a new US or Canadian number, transfers to team members ('some limitations apply'), missed-call messages by email, 1 team member, 1,000 characters of business knowledge, basic reporting.
Premium, $20 a month: your own number, messages by email or text, 10,000 characters, 10 team members, advanced reporting with optional recordings and transcriptions, more voices.
Pro, $189 a month, 'coming soon' with a waitlist: 20 team members, warm transfer, two-way texting, recording, transcription and 'team scheduling'. Enterprise is by request. A yearly toggle shows 20% off.
No plan lists minutes, call limits or an extra-minute price, and none lists booking. 200 calls fit on Free at $0 on the page's own wording, or on Premium at $20 for recordings and texts.
- 'Unlimited usage' is the page's wording; no minute or call limit is stated, and I did not run 200 calls. The Premium and Pro prices were not purchased.
Other features checked
| Text messages | Not Tested | Texted messages are a Premium feature, and Pro adds two-way texting. I used the Free plan, so no text was sent. |
| Recordings and transcripts | Not Tested | Premium lists optional recordings and transcriptions; Free has basic call reporting. I did not use Premium. |
| Mobile app | Unavailable | Nucleus's FAQ page, read on 24 October 2026, says 'No, currently there is no Nucleus phone app.' The web dashboard is the only way in, and I used it; messages come by email. |
| Calendar connections | Not Tested | No booking, scheduling or calendar is listed on Free, Premium or Pro; Pro, 'coming soon', lists 'team scheduling' only. |
| Own number or forwarding | Not Tested | Free gives a new US or Canadian number; bringing your own number is a Premium feature. I used the new number. |
Coverage
Tested
- Six graded calls to the agent's phone number, plus a second take of the call for the owner
- 11 business facts stated on the graded calls, each checked against the price list and FAQ
- Three message calls checked against the emails Nucleus sent
- A transfer to the owner, twice, and an emergency call that should have been transferred
- 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 pricing page on 24 October 2026
Not tested or limited
- The voice is not graded: I rate voices only by ear, and these recordings have not been rated yet
- Setup is not timed, so it is not graded
- Booking is not graded: Nucleus lists no booking on any plan, so the booking, move and cancellation calls were not run
- There is no recording of call 5 (the call for the owner), on either take, so the silence after 'being transferred' cannot be heard; its timings come from the call log
- Dan's number in the staff directory was the Google Voice line every test call came from, so a transfer to him could only reach a busy line; the result shows what Nucleus does when a transfer cannot connect, not a transfer that completes
- Nucleus called my fictional number, five five five, five five five, five five five five, incomplete once on call 3 and twice on call 6; a real number would not have caused that loop, so the messages grade does not rest on it
- All calls ran in the firm's Saturday opening hours; Nucleus has no opening-hours setting on the Free plan, only each staff member's availability, so no true after-hours call was run
- The Free plan lists one team member; both of mine were saved and put live, so I did not see that limit enforced
- The pricing page was read on 24 October 2026; I did not open checkout for Premium
Still to test
- The voice, rated by ear or by a panel of listeners
- A transfer to a number a person answers, to hear what the receptionist and the person say, with a recording
- A true after-hours call with Dan switched unavailable
- Setup timed end to end, and the Premium plan: recordings, texts and a forwarded business line
- The senior discount, the warranty and the payment methods, asked directly