AI Receptionist for Small Business: Cost and ROI

If you run a small business, you already know the feeling: the phone rings while you're under a sink, with a client, or driving between jobs, and by the time you call back, the customer has booked someone else. An AI receptionist for small business answers every call, books appointments and sends you a summary, for a fraction of the cost of a person at the front desk. Whether it pays for itself depends on three numbers you can estimate this afternoon: how many calls you miss, how many of those are new customers, and what a new job is worth to you. This guide walks through the math, the options and the trade-offs.

What missed calls really cost

Start with how often it happens. In a study published in January 2016, 411 Locals tracked calls to 85 businesses across 58 industries for 30 days. On average, 37.8% of calls were answered by a person, 37.8% went to voicemail and 24.3% got no response at all. Seventy percent of the businesses answered fewer than half of their calls. The study is small and a decade old, so treat it as a sign of the problem rather than a benchmark for your business. The best number is always your own: most phone systems and carriers can show your missed-call count for the last 30 days.

Then consider what happens after the missed call. In a Harvard Business Review study, researchers audited 2,241 U.S. companies by sending a web inquiry and timing the reply. Only 37% responded within an hour, and 23% never responded at all. A separate analysis of 1.25 million sales leads found that firms trying to contact a lead within an hour were nearly seven times as likely to have a meaningful conversation with a decision maker as firms that tried even an hour later, and more than 60 times as likely as firms that waited 24 hours or longer. That research was on web leads, not phone calls, but the lesson is the same: interest goes cold fast.

A voicemail is not a booking. Many callers don't leave one, and the ones who do may keep dialing your competitors while they wait.

What an AI receptionist does (and doesn't)

A modern AI receptionist answers the phone in a natural voice, using your business name and the information you give it. Typical features include:

• Answering questions about hours, service area, pricing ranges and services.
• Booking, rescheduling or canceling appointments in your calendar.
• Taking a message and sending you a summary by text or email.
• Transferring urgent calls to you or a staff member.
• Handling text messages as well as calls.

What it shouldn't do:

• Quote a firm price for complex work it can't see.
• Handle emergencies that need judgment, such as a gas leak or a medical issue. It should route those to a person or tell callers to contact emergency services.
• Pretend to be a human if someone asks. Be upfront.

Your options, side by side

The BLS reports that the median hourly wage for receptionists was $18.27 in May 2025, or about $38,010 a year, and that's before payroll taxes and benefits. A person also can't answer two calls at once, and nobody is at the desk at 9 p.m. when a pipe bursts.

Worked example: does it pay back?

Here's a way to run the numbers for your own business. The example is a small plumbing company, but the steps work for any appointment or service business.

Step 1: Count missed calls. Pull 30 days from your phone system. Say it shows 60 missed calls.

Step 2: Estimate how many were new customers. Some are existing customers, suppliers or spam. Listen to a sample of voicemails, or check the numbers against your customer list. Say one-third, or 20 calls, were potential new jobs.

Step 3: Estimate how many would have booked. Not every caller books even when you answer. Use your own close rate on answered calls. Say 25%, so 5 jobs a month.

Step 4: Estimate the gross profit per job. Revenue minus materials and labor. Say $250.

Step 5: Multiply. 5 jobs × $250 = $1,250 a month in gross profit that an always-on receptionist could recover.

Illustration, not a prediction: at $1,250 a month recovered, any subscription that costs less than that each month covers itself, and a $12,600 owned app pays back in about 10 months before running costs. Your numbers depend on your call volume, close rate and pricing.

If your numbers come out small, say two missed new-customer calls a month, a simple voicemail-to-text setup and a callback habit may be all you need. That's a fine answer.

How to choose and set one up

Test with real calls. Before committing, call the AI yourself 10 times with realistic questions, including awkward ones. Ask about something you don't offer. See whether it admits it doesn't know.

Write a clear knowledge sheet. List services, service area, hours, price ranges, what counts as an emergency, and when to transfer. The AI is only as good as what you give it.

Connect it to your calendar. Booking directly is where the value is. A message-taker alone recovers fewer jobs.

Decide on disclosure. Many owners have the receptionist say it's an automated assistant at the start. Being clear up front is the honest approach, and it sets expectations if the AI gets something wrong.

Know the rules on recording and outbound calls. Recording-consent laws vary by state, and some require every party to agree, so check yours. For outbound calls, the FCC ruled in February 2024 that AI-generated voices count as "artificial" under the Telephone Consumer Protection Act, which means robocalls using them need the recipient's prior express consent. Answering inbound calls is different, but an AI that places marketing callbacks is covered.

Review summaries weekly for the first month. Fix wrong answers in the knowledge sheet. Look for patterns in what callers ask.

When an off-the-shelf tool is the better choice

We'll be direct: for most single-location businesses, a monthly AI answering subscription from an established provider is the right first move. It's quick to set up, costs little to try, and you can cancel if it doesn't work. Try one for a month and measure booked jobs against your baseline.

Owning a branded AI receptionist app makes more sense when:

• You run several locations or brands and want one system you control.
• You need it to fit your own booking rules or software in ways subscription tools don't allow.
• You want to offer it to other businesses in your trade, for example as an association or a franchise group, and run it as a product.

If none of those apply, rent, don't build. Our breakdown of whether an app pays for itself covers the ownership math in more depth.

Missed calls are usually part of a bigger follow-up problem. Once calls are answered, the next leak is slow responses to web and text leads, which we cover in automated lead follow-up. For appointment businesses, the leak after that is no-shows.

How Ottomind can help

The Ottomind Store sells an AI Receptionist for Small Businesses app that we build to order and brand for you. It answers calls and texts in your business's voice, books appointments into your calendar and sends lead and call summaries to your inbox. You can own 100% for $18,000, or own 80% for $12,600, with Ottomind keeping a 20% non-voting share that you can buy back within 24 months for the price difference. The 80% option can be paid as $500 monthly installments with no interest. It's built in about four weeks after the first payment, and iOS and Android versions are quoted add-ons.

If you'd rather not pay the price upfront, the Starter lets you start a Store web app at $500 a month, for up to 12 months. It's a rental: payments don't count toward the price, Ottomind owns the app in the meantime (hosting and upkeep included), and you can buy it any time. If you don't buy it after 12 months, it's switched off. And if a low-cost monthly subscription covers what you need, use that instead.