How Many AI Receptionist Minutes Does a Locksmith Shop Need? A Sizing Guide From Your Own Call Log (2026)

How Many AI Receptionist Minutes Does a Locksmith Shop Need? A Sizing Guide From Your Own Call Log (2026)

You have decided to try an AI receptionist. You open the pricing page and find plans measured in minutes: 500, 1,000, 2,500. You know roughly how busy the phone is. You do not know how many minutes that is.

Most owners guess. Some guess high and pay for minutes they never use. Some guess low, run over in the second week, and decide the product is expensive when the real problem was the plan.

As of October 2026, there is no need to guess, because the number you need is sitting in your own phone records. Sizing an AI receptionist plan is one multiplication: calls per month times average handled minutes per call. This guide shows where to find both numbers, how to do the multiplication honestly, which TheKeyBot plan the result maps to, what overage costs at each plan's rate, when the 100 calls included in KeyBot Lite is enough, and how to check your estimate after the first month.

Two notes before starting. Every shop example below is hypothetical and labelled as such; they exist to show the arithmetic, not to tell you what a typical shop looks like. And the only prices used are TheKeyBot's own published plan numbers. Your call volume is yours to measure.

Calls and minutes are two different units

TheKeyBot sells two kinds of product, and they are metered differently.

KeyBot Lite is a message-taking receptionist. It answers, takes the caller's details and sends the message to your team through Telegram. It does not quote or book. It costs $149 a month, includes 100 calls a month, and is 50 cents a minute after that. The first 5 answered calls are free on a 7-day trial.

The full platform plans quote automotive keys by year, make and model, book appointments, dispatch the nearest technician and send payment links. They are metered in AI minutes:

• Core: $500 a month, 500 AI minutes, 45 cents a minute after
• Pro: $750 a month, 1,000 AI minutes, 40 cents a minute after
• Elite: $1,200 a month, 2,500 AI minutes, 35 cents a minute after

Each full plan has a 14-day free trial, unlimited users and no per-seat fees.

So the sizing question has two halves. For Lite, you need to know how many calls you get. For the full plans, you need calls and the average length of a call. Both start with the same piece of homework.

A longer walk through what each plan includes is in the guide to TheKeyBot pricing explained. This guide is only about getting the size right.

Step 1: Count the calls in your own log

Your phone carrier, your VoIP system or your call tracking tool keeps a record of every inbound call: the date, the time, the number, whether it was answered, and how long it lasted. Export one full month. If you can, export three and use them all.

Then count carefully. The number you want is calls the AI receptionist would answer, which is not the same as total inbound calls.

Decide what the receptionist will cover. There are three common setups, and they produce very different counts:

• Everything. Every inbound call goes to the AI first. Count all inbound calls.
• Overflow and after-hours. The AI answers only when your staff do not pick up within a few rings, and outside business hours. Count missed calls, calls that went to voicemail, and calls outside your staffed hours.
• After-hours only. Count only calls outside staffed hours.

A shop that sizes for "everything" and then deploys "after-hours only" has bought far too much. Decide the setup first.

Include the calls you missed. Calls that rang out or reached voicemail are exactly the calls an AI receptionist exists to answer. They belong in the count even though nobody handled them last month. If you have never looked at how many there are, that number alone is worth the exercise, and the missed call cost calculator will help you put your own figures to it.

Leave out outbound calls and internal calls. Calls your technicians make to customers are not inbound.

Think about spam separately. Robocalls and sales calls will reach an AI receptionist just as they reach you. Most are short. Count them, because they are calls, but keep them in their own column so you can see how much of your volume they are. Background on unwanted calls is published by the Federal Communications Commission, and how screening works on a shop line is covered in the guide to spam calls and robocall screening.

If you have never organized a call log before, the guide to the seven numbers in a locksmith call log sets out what to pull and how to read it.

At the end of this step you have one number: calls per month the receptionist would answer.

Step 2: Find your average handled minutes

The second number is how long a call lasts when it is actually handled. Your log has a duration column. Use it, with three cautions.

Use answered calls only. A missed call has a duration of zero or of a few seconds of ringing. Including those drags the average down and makes you under-buy. Filter to calls that were answered and lasted long enough to be a conversation.

Use the mean, and look at the spread. Add up the durations of all handled calls and divide by the number of calls. Then look at the longest and shortest. A locksmith line has very short calls (wrong numbers, "are you open?", spam) and long ones (a caller who does not know what vehicle they have, a commercial customer describing a rekey). The mean covers both, which is what you want for billing.

Expect the AI's calls to differ a little from yours. An AI receptionist on a full plan asks a set list of questions, quotes and books. That is often a longer call than an owner saying "I'll call you back in ten minutes," and a shorter call than a chatty dispatcher. A message-taking call on Lite is usually shorter than a quote-and-book call. There is no universal figure to give you here, and anyone who offers one is guessing. Your own log is the starting estimate, and the first month on the product is the real measurement.

If your phone system gives you no durations at all, time twenty real calls with a stopwatch across a normal week, including a few after-hours calls. Twenty is not a study. It is far better than a guess.

At the end of this step you have the second number: average handled minutes per call.

Step 3: Multiply

Monthly AI minutes = calls per month × average handled minutes.

Here is a worked example with a hypothetical shop. The shop and its numbers are invented to show the arithmetic.

A hypothetical shop's log shows 300 calls a month that the receptionist would answer, and an average handled call of 2.5 minutes.

• 300 calls × 2.5 minutes = 750 minutes a month

That is the number to take to the plan list.

Step 4: Map the minutes to a plan

For any monthly total, the cost on a plan is the monthly price, plus any minutes over the included amount multiplied by that plan's overage rate.

Take the hypothetical 750 minutes:

• Core: 500 minutes are included. 750 − 500 = 250 minutes over. 250 × $0.45 = $112.50. Total: $500 + $112.50 = $612.50
• Pro: 1,000 minutes are included, so 750 is covered with 250 to spare. Total: $750
• Elite: covered with room to spare. Total: $1,200

Two honest readings of that result. Pro at $750 covers all 750 minutes inside its included 1,000 and leaves headroom for a busy stretch. Core is cheaper on these exact numbers, at $612.50, because paying for 250 overage minutes costs less than the $250 step up to Pro. Which is right depends on how steady your volume is, which is covered below.

The break-even points

Because the plans have different overage rates, there are two volumes where the cheaper plan changes. The arithmetic:

Core and Pro cost the same at 1,500 minutes. Core: $500 + (1,000 minutes over × $0.45) = $950. Pro: $750 + (500 minutes over × $0.40) = $950. Below 1,500 minutes a month, Core with overage is the lower bill. Above it, Pro is.

Pro and Elite cost the same at 2,125 minutes. Pro: $750 + (1,125 minutes over × $0.40) = $1,200. Elite: $1,200 with 2,125 inside its 2,500 included minutes. Below 2,125 minutes, Pro with overage is the lower bill. Above it, Elite is.

So, on price alone:

• Up to 1,500 minutes a month: Core
• From 1,500 to 2,125 minutes: Pro
• Above 2,125 minutes: Elite

Price alone is not the whole decision. A plan with more included minutes gives a predictable bill, and some owners value that over saving a few dollars in an average month. Check the pricing page for anything else that differs between plans before choosing on minutes alone.

Monthly cost by volume, at a glance

Every row is arithmetic on the published plan prices. The minute totals are hypothetical volumes chosen to show where the lines cross.

To check any row yourself: subtract the included minutes from the total, multiply what is left by the overage rate, and add the monthly price. For example, 3,000 minutes on Elite is 500 minutes over, 500 × $0.35 = $175, and $1,200 + $175 = $1,375.

Four hypothetical shops

These are invented examples. None is a real customer and none is offered as typical.

Hypothetical shop A: one truck, after-hours messages only. The owner answers during the day and wants the phone covered at night and on weekends. The log shows 80 after-hours calls a month. The owner does not want quoting or booking, only a clean message with the caller's details. That is under the 100 calls included in KeyBot Lite. Cost: $149 a month.

Hypothetical shop B: two trucks, overflow and after-hours. 180 calls a month reach the receptionist, at an average of 2 minutes. 180 × 2 = 360 minutes. That fits inside Core's 500 minutes with 140 to spare. Cost: $500 a month.

Hypothetical shop C: three trucks, every call answered by the AI. 700 calls a month at 2.5 minutes. 700 × 2.5 = 1,750 minutes.

• Core: 1,250 over × $0.45 = $562.50, total $1,062.50
• Pro: 750 over × $0.40 = $300, total $1,050
• Elite: covered, total $1,200

Pro is the lowest bill, by a small margin over Core.

Hypothetical shop D: a larger shop running ads. 1,100 calls a month at 2.5 minutes. 1,100 × 2.5 = 2,750 minutes.

• Core: 2,250 over × $0.45 = $1,012.50, total $1,512.50
• Pro: 1,750 over × $0.40 = $700, total $1,450
• Elite: 250 over × $0.35 = $87.50, total $1,287.50

Elite is the lowest bill.

A fuller version of this kind of reasoning for a small mobile operation is in the guide to AI receptionist ROI for a two-truck mobile locksmith.

When KeyBot Lite's 100 calls is enough

Lite is the right size when two things are true.

You want messages, not quotes and bookings. Lite takes the call, collects the details and sends them to Telegram. A person on your team calls the customer back to quote and schedule. If you upload a price sheet and confirm it on screen, Lite states those exact prices when asked and does not negotiate or invent one, but it does not run a quoting engine or book the job. If you want the call closed without a callback, that is the full platform. What Lite does and does not do is laid out in the guide to KeyBot Lite as a message-taking receptionist.

The calls it will answer are around 100 a month or fewer. For a one-truck shop covering nights and weekends, or a shop that uses it only for overflow, that is often the case. Count from your log, not from memory.

If you go over, Lite is 50 cents a minute after the included 100 calls. A hypothetical example: a shop receives 130 calls in a month, so 30 calls fall beyond the included 100, and those calls average 2 minutes each. 30 × 2 = 60 minutes. 60 × $0.50 = $30. The month costs $149 + $30 = $179. That example assumes the per-minute rate applies to the minutes on calls beyond the first 100; confirm exactly how overage is counted on the pricing page before you rely on it.

Lite stops being the right size when the overage becomes routine, or when the callbacks themselves become the bottleneck: the messages arrive, and nobody has time to return them before the customer calls another shop. That second sign matters more than the bill. The decision is covered in the guide to when to upgrade from Lite to Core.

For most shops that have never used an AI receptionist, KeyBot Lite is the place to start, precisely because the first month on it produces the call count and call lengths this guide asks you to estimate.

What pushes minutes up or down

Your estimate will move for reasons that have nothing to do with how many customers you have.

Seasonal swings. A cold snap, a holiday weekend or the start of a school term can lift lockout calls well above an average month. If you sized on a quiet month, you will run over in a busy one. Use three months of logs if you can, and note which was the busiest. Planning for those spikes is covered in the guide to seasonal call volume spikes.

Advertising. If you are about to turn on or increase ads, last month's log understates next month's calls. This matters most for pay-per-lead channels, where every unanswered call is money spent for nothing; the answer to the question are Local Services Ads worth it depends heavily on your answer rate. Google's own description of the program is on the Local Services help pages.

Spam. A line that attracts a lot of robocalls will see a lot of short calls. They add little per call and can add up.

The intake you configure. A receptionist that asks twelve questions runs longer calls than one that asks six. More questions mean better tickets and more minutes. That is a trade to make on purpose.

Bilingual callers. Calls are answered in English and Spanish. A shop that was previously losing Spanish-speaking callers at "hello" may see handled calls rise simply because those callers now stay on the line.

Repeat callers. A customer who calls three times to ask where the technician is counts three times. Better status texts reduce those calls.

Headroom: steady shop or spiky shop

Once you have a monthly estimate, ask how much it moves from month to month.

If your volume is steady, choose on the break-even points above. Overage on Core up to 1,500 minutes is the lower bill, and you will rarely be surprised.

If your volume is spiky, look at your busiest month, not your average. A hypothetical shop averaging 1,400 minutes with a peak month of 2,000 would pay $905 on Core in an average month (900 over × $0.45 = $405, plus $500) and $1,175 in the peak month (1,500 over × $0.45 = $675, plus $500). On Pro it would pay $910 in the average month (400 over × $0.40 = $160, plus $750) and $1,150 in the peak (1,000 over × $0.40 = $400, plus $750). For that shop the two plans come out within a few dollars of each other, and the choice is about which bill you would rather see.

There is no penalty for being wrong in the first month, as long as you look.

Re-check after the first month

An estimate from a call log is a starting point. After the first full month on the product you have something better: the real numbers. Do this review once, at the end of month one, and again each quarter.

1. Actual calls answered. Compare with your estimate from Step 1. If it is much higher, you were missing more calls than your log showed.
2. Actual minutes used. Compare with Step 3.
3. Actual average minutes per call. Divide minutes by calls. This replaces your estimate from Step 2 for good.
4. Overage paid. If you paid overage, was it a one-off or the new normal?
5. Where you sit against the break-evens. Consistently over 1,500 minutes on Core, move to Pro. Consistently over 2,125 on Pro, move to Elite. Consistently far under your included minutes, move down.
6. What the calls were. Sort by type. If a third of your minutes are spam or wrong numbers, that is a screening question, not a plan question.

On Lite, the review is simpler: how many calls, how many beyond 100, and how quickly the messages were returned.

The trials exist for this. The full plans have a 14-day free trial and Lite's first 5 answered calls are free on a 7-day trial. Use that window to measure, not only to listen.

Comparing against other options

Owners sizing an AI plan are usually also weighing a human answering service or an in-house hire. A few pointers, without claims about any particular company.

Per-minute answering services are sized the same way as this guide, with one extra step: find out how the service counts a minute. Rounding rules, minimum call lengths and what counts as billable time vary by contract and change the effective rate. The questions to ask are in the guides to per-minute versus flat-rate answering service pricing and answering service contract terms. If you are comparing vendors by name, the roundups of AnswerForce alternatives and alternatives to hiring a receptionist describe what to look for; verify every price on the vendor's own page, because they change.

An in-house receptionist is not sized in minutes at all. The cost is wages, payroll costs and coverage for the hours that person is not there. Occupational wage data for your area is published by the Bureau of Labor Statistics, and it is the right place to get a figure rather than a blog post.

Ask every vendor the same four questions: what is a billable minute, what is the overage rate, what happens in a month when I go far over, and can I change plans without a penalty.

Common sizing mistakes

• Guessing volume from memory instead of exporting the log.
• Sizing on answered calls only, and leaving out the missed calls the receptionist will now pick up.
• Averaging in zero-length missed calls, which makes the average call look shorter than it is.
• Sizing for "every call" and deploying "after-hours only," or the reverse.
• Using a slow month as the baseline.
• Turning on ads the same week without adding to the estimate.
• Assuming the bigger plan is always the safer buy. Below 1,500 minutes, Core with overage is the lower bill.
• Never re-checking. The first month's real numbers are worth more than any estimate.

What it costs

KeyBot Lite is $149 a month with 100 calls a month included and 50 cents a minute after that, and the first 5 answered calls are free on a 7-day trial. It is message-taking only and delivers messages to Telegram.

The full platform plans are Core at $500 a month with 500 AI minutes and 45 cents a minute after, Pro at $750 a month with 1,000 minutes and 40 cents a minute after, and Elite at $1,200 a month with 2,500 minutes and 35 cents a minute after. Each has a 14-day free trial, unlimited users and no per-seat fees. Details are on the pricing page.

To hear a call before doing any of this arithmetic, the instant locksmith demo has the AI call your phone in about 30 seconds, and the live demo line is +1 (716) 350-6391.

The bottom line

The number of AI receptionist minutes a locksmith shop needs is its calls per month multiplied by its average handled minutes per call, and both figures come from the shop's own call log. Decide first what the receptionist will cover, count the calls it would answer including the ones you currently miss, and take the average length of answered calls only. Map the result to a plan by adding the monthly price to any overage: on price alone, Core is the lower bill up to 1,500 minutes a month, Pro from 1,500 to 2,125, and Elite above that. If you want messages rather than quoting and booking, and the receptionist would answer around 100 calls a month or fewer, KeyBot Lite at $149 is the right size. Add headroom if your volume is seasonal or you are about to advertise. Then check the real numbers at the end of the first month and move up or down. The estimate gets you started; the first invoice tells you the truth.

Frequently asked questions

How do I estimate how many AI receptionist minutes my locksmith shop needs?

You estimate AI receptionist minutes by multiplying the number of calls the receptionist would answer each month by your average handled minutes per call. Both numbers come from your own phone records. Count the calls for the coverage you plan to use, include the calls you currently miss, and average the length of answered calls only. As a hypothetical example, 300 calls at 2.5 minutes each is 750 minutes a month.

Which TheKeyBot plan fits a given number of minutes?

On price alone, Core is the lowest bill up to 1,500 minutes a month, Pro is lowest from 1,500 to 2,125 minutes, and Elite is lowest above 2,125 minutes. Those break-even points come from the published prices: Core is $500 with 500 minutes and 45 cents a minute after, Pro is $750 with 1,000 minutes and 40 cents after, and Elite is $1,200 with 2,500 minutes and 35 cents after.

What does overage cost if I go over my plan's minutes?

Overage costs 45 cents a minute on Core, 40 cents a minute on Pro and 35 cents a minute on Elite, charged only on minutes beyond the plan's included amount. As a hypothetical example, a shop using 750 minutes on Core is 250 minutes over, and 250 times 45 cents is $112.50, for a monthly total of $612.50. The same 750 minutes fits inside Pro's 1,000 included minutes at $750.

Is KeyBot Lite's 100 calls a month enough for a locksmith shop?

KeyBot Lite's 100 included calls is enough when the receptionist would answer around 100 calls a month or fewer and you want messages taken rather than jobs quoted and booked. That often describes a one-truck shop covering nights and weekends or a shop using it for overflow. Lite costs $149 a month and is 50 cents a minute after the included calls, so count the calls in your own log before deciding.

How much do TheKeyBot plans cost?

KeyBot Lite costs $149 per month with 100 calls included and 50 cents a minute after that, with the first 5 answered calls free on a 7-day trial. The full platform plans are Core at $500 per month with 500 AI minutes and 45 cents a minute after, Pro at $750 with 1,000 minutes and 40 cents after, and Elite at $1,200 with 2,500 minutes and 35 cents after, each with a 14-day free trial. Details are at https://www.thekeybot.com/pricing.

When should I re-check whether I am on the right plan?

You should re-check at the end of the first full month and then once a quarter, using the actual calls answered, minutes used and overage paid. Divide minutes by calls to get your real average call length, which replaces your original estimate. If you are consistently above 1,500 minutes on Core or above 2,125 on Pro, move up a plan, and if you are consistently far below your included minutes, move down.