Is an AI Receptionist Worth It? Work Out Your Own Number

Nobody can answer this for you, and anybody who tries without asking about your business is selling.

“Worth it” is not a property of the product. It is a relationship between what you are losing and what the thing costs. Two contractors can buy the identical system at the identical price and one of them made a great decision while the other wasted the money. The difference is entirely in numbers they already have.

Here is how to get yours.

The three numbers that decide it

Everything else is noise. You need these, and you can get all three in an afternoon.

One: how many inbound calls you did not connect last month. Not how many you think. Pull the log from your carrier. Every major US carrier gives you inbound call records in the online account, and it takes about twenty minutes to export a month and count the ones that never connected. Do not estimate this. Owners are almost always wrong about it in both directions.

Two: what a job is worth to you on average. Revenue, not profit, for this calculation. Take last year’s total revenue and divide by the number of jobs. Do not use your best job. Do not use the one you tell people about.

Three: your close rate on calls you actually answer. Of the calls that reach a human, what share become work? Most contractors have never measured this and are surprised by it. If you genuinely cannot get it, use a conservative guess and be honest that it is a guess.

The arithmetic

The following numbers are made up to show the shape of the calculation. Use your own. A number you plug in from an article is worse than no number at all.

Missed calls × close rate × average job value = revenue that walked

Say a waterproofing company missed 38 inbound calls last month. Their close rate on answered calls is 30%. Average job is $4,200.

38 × 0.30 × $4,200 = $47,880

Now apply the honesty discount, because that figure is too high and you should not believe it.

Three corrections that most versions of this skip

Not every missed call was a job. Suppliers, wrong numbers, spam, your own crew, people outside your service area, the same person calling three times in ten minutes. In most call logs a real chunk of the misses were never going to be revenue. Go through a sample of fifty by hand and work out what share were genuine prospects. Apply that percentage.

Some of them called back. A person who really wants you specifically will try again. Check your log for repeat numbers. Those were not lost, just delayed.

You could not have serviced all of them anyway. If your crews were booked solid, the tenth emergency call that day was going somewhere else regardless. Capacity is a real ceiling and it does not care what answers your phone.

Run those three corrections and the $47,880 in the example above might land somewhere near a third of that. That is still a large number against a subscription. But it is a number you can defend, and the point of doing this properly is that you are the one who has to live with the decision.

The break-even question

Once you have a corrected monthly figure, the question becomes simple.

Divide the monthly cost of the system by your average job value. That tells you how many extra booked jobs a month it has to produce to pay for itself. For most contractors in these trades it lands somewhere between a fraction of one job and two jobs a month, because the ticket sizes are large relative to the subscription. Pricing varies and nobody publishes a straight number, which is a whole subject of its own in what an AI receptionist costs.

If the answer is “it needs to book one extra job a month,” and you missed 38 calls, the decision is not difficult. If the answer is “it needs to book six,” be much more careful.

When the honest answer is no

You are already at capacity and turning work away. This is the big one. If your constraint is crews, not calls, answering the phone better produces a longer list of people you are letting down. Every extra booked job you cannot service costs you a review and a referral. Fix throughput first. This is not a small caveat, it is the most common reason the purchase disappoints.

Your close rate on answered calls is poor. If you answer the phone and still lose most of them, the problem is downstream of the ring. More conversations will mostly produce more losses. Work out why they are not converting before you buy more of them.

Your volume is genuinely low. A handful of calls a week that you almost always answer means there is very little to recover. The system works fine. There is just nothing for it to do.

Almost all your work is commercial or repeat. Those callers wait for you. Cold search traffic does not. If you have no cold inbound to speak of, this is solving somebody else’s problem.

Your work arrives through insurance programmes or a TPA. Your volume is set by your position on somebody’s list, not by your phone. Answering faster is nice. It does not move the number.

When it is close to obvious

The mirror image: seasonal or weather-driven demand, where your best days are also the days nobody can pick up. Storm-driven waterproofing, freeze-driven restoration, the compressed concrete season. On those days a human office cannot scale and an answering service bills you more precisely when you are busiest. That asymmetry is the strongest version of the case, and it is trade-specific rather than general.

Large average job value helps for the obvious reason. When one booked job covers several months of subscription, the break-even arrives fast and the risk of the decision is small.

Before you sign anything

Call the vendor’s demo line and behave like your worst customer. Interrupt it. Change your mind mid-sentence. Ask something outside its scope and see whether it admits it or invents an answer. The demo you get walked through is not the test. The test is whether it survives a caller who is not cooperating.

Then ask what happens on the calls it cannot handle, because those exist and any vendor claiming otherwise is worth walking away from. The rest of the list is in what to ask before you buy an AI voice agent.

If you have run your own numbers and the gap is real, that is when it is worth talking to somebody about AI voice agents for contractors. If you have not run them yet, do that first. It costs an afternoon and it is the only part of this decision nobody can do for you.