AI Receptionist ROI: How to Measure It Honestly
AI receptionist ROI is the value of the calls you stop missing minus what the service costs. Multiply your missed calls per month by the share you would have closed and the average job value, then subtract the monthly fee. Use your own numbers, not a vendor's averages, and track missed calls for two weeks before deciding.
Why most AI receptionist ROI claims are useless
Search for this topic and you will find calculators that already know your answer. They plug in an industry average for missed calls, an average close rate, an average ticket, and hand you a large number with a buy button under it. None of those averages are yours. A one-truck plumber and a six-crew roofing company do not miss the same calls, close at the same rate, or invoice the same amounts.
The honest version is boring. It is a formula with four or five inputs, and every input comes from your own phone log and your own invoices. This post gives you the formula, tells you where each number lives, and walks through it with placeholder letters so you can substitute your own figures. If you want the calculator version, our ROI page runs the same math.
The formula, with your variables
Start with what a missed call is worth to you, then compare that with what coverage costs. Two lines are all you need.
Recovered value per month = M × C × J. Net gain per month = (M × C × J) minus (F + R × T).
Here is what each letter means. M is the number of calls per month you currently miss that were real job requests, not spam or wrong numbers. C is your close rate on a callable lead, written as a decimal. J is the value of an average job, ideally gross margin rather than revenue. F is the flat monthly fee of the service. R is the per-minute rate, if the service has one. T is the minutes of calls the agent handles per month.
For Trexinet, F is 59 and R is 0.05, so the cost side reads 59 + 0.05 × T. There is no contract, so the cost line stops the month you stop.
Where to find each number
Every input is available to you today without buying anything. The table shows where to look and the mistake people most often make with each one.
| Variable | What it means | Where to find it | Common mistake |
|---|---|---|---|
| M, missed calls | Real job requests per month that went unanswered | Carrier call log: ring-no-answer and busy, plus voicemails and after-hours calls | Counting every missed call, including spam and vendors |
| C, close rate | Share of answered job requests that became booked work | Last quarter: booked jobs divided by qualified calls answered | Using the close rate on quotes instead of on calls |
| J, job value | What an average job is worth to you | Invoices from the same quarter, margin if you know it | Using your biggest jobs instead of the typical one |
| F, flat fee | The monthly base price of the service | The provider's pricing page | Ignoring setup fees or minimum terms |
| R × T, usage | Per-minute rate times minutes handled | Average call length from your phone log times expected calls | Forgetting that the agent also answers calls you would have answered |
A worked example using placeholders
Suppose your two-week count of missed job calls, doubled to a month, gives you M. Your booked-to-answered ratio from last quarter gives you C. Your invoices give you J. Multiply them: M × C × J is the monthly value you are currently leaving on the table.
Now the cost side. Your phone log tells you the average call runs a certain number of minutes. Multiply that by the calls you expect the agent to take, both the ones you miss now and the ones it will pick up during business hours, and you have T. Cost is F + R × T. With Trexinet that is 59 + 0.05 × T.
Break-even is the point where M × C × J equals F + R × T. Rearranged, you need at least (F + R × T) divided by (C × J) recovered calls per month to cover the service. Compare that number with your M. If M is comfortably above it, the service pays for itself on recovered calls alone. If M is below it, the service is not an ROI decision, it is a convenience decision, and you should treat it as one.
Run the calculation twice. Once with your optimistic numbers and once with your pessimistic ones. If both versions clear break-even, you have your answer. If only the optimistic one does, keep counting for another two weeks.
What the formula leaves out
The two-line formula is deliberately narrow. It only counts recovered first-time bookings. Real life adds items to both sides.
- On the benefit side: repeat customers from jobs you would have missed, reviews from callers who got picked up at 9 p.m., and the hours you or your office staff stop spending on callbacks.
- On the benefit side: calls that get booked during the day while your team is on a job, which are not missed calls in the log but are often callbacks that never connect.
- On the cost side: your time reviewing summaries in the first weeks, the occasional booking that lands in the wrong slot, and any caller the agent handles worse than you would have.
- On the cost side: setup effort. A managed provider absorbs most of it. A self-serve tool puts it on you.
How to measure it honestly after you start
The mistake most owners make is measuring nothing before and everything after. Fix that with a baseline. Count missed job calls for two weeks before the agent goes live. Keep the same count running after.
Then track three things from the agent's own summaries: calls answered outside your office hours, bookings made, and jobs completed from those bookings. The last one is the only number that matters. A booking that no-shows is not revenue. Compare completed jobs from agent bookings against the cost line each month.
Give it one full billing cycle plus your typical gap between booking and invoice. For a same-week trade, that is about six weeks. For a roofing or remodel business with long quotes, it is longer. Judge it at that point, not at day ten.
What Trexinet costs, so you can finish the math
$59 per month plus $0.05 per minute, no contract. Setup is done by Trexinet on your real calls, the agent answers 24/7, books on your calendar, and texts a summary after every call. Put 59 and 0.05 into the cost line above, use your own M, C, and J, and you have an honest number. The ROI page does the arithmetic for you if you would rather not.
Related reading
Frequently asked questions
What close rate should I use?
Your own, from a recent quarter: booked jobs divided by qualified calls you actually answered. Do not use your close rate on written quotes, which is a different funnel stage. If you have never tracked it, tally answered job calls and bookings for two weeks and use that. It will be rough, but it will be yours.
Should I use revenue or margin for job value?
Margin is more honest. A recovered job that pays for parts and a technician's day is worth the profit on it, not the invoice total. If you do not know your margin per job, use revenue but be aware the result is inflated, and run the pessimistic version of the formula as well.
How do I count missed calls without special software?
Pull the call log from your carrier or phone system for two weeks. Count ring-no-answer, busy, and after-hours calls. Listen to voicemails and remove spam, vendors, and existing customers who reached you another way. What is left is M for the half-month. Double it for a monthly figure.
Does an AI receptionist pay off for a low-volume business?
Only if your recovered value clears the cost line. With a low flat fee, the break-even count of recovered calls is small, but if you miss almost nothing today, there is nothing to recover. In that case treat it as convenience, not ROI, and decide on that basis. The formula will tell you which case you are in.
How long before I know whether it worked?
One full billing cycle plus your usual gap between a booking and an invoice. That gives you completed jobs to count, not just bookings. Judge on completed jobs from agent bookings against the fee, and compare with the missed-call baseline you counted before you started.