How does AI customer service drive revenue for a local business?

Where the revenue comes from

An AI assistant does not create demand. It catches demand you already have and would otherwise miss. For a local business, look in three places.

• Answers outside your hours. Zendesk's CX Trends 2026 research, a survey of 6,182 consumers in 22 countries in June 2025, found that 74% of consumers now expect 24/7 service because of AI. A question answered at 11pm can still become a sale tomorrow.
• Orders and carts built in the chat. A customer who asks for two burgers and a shake can get the order read back with the total and a link to pay, instead of a reply the next morning.
• Leads captured while you work. A customer who asks about a repair can leave the job, the address and a phone number for you to call back.

Orders: from a question to a checkout

For a restaurant, the revenue moment is the order. An assistant that answers from your real menu and prices can take the order, ask for any required choice, and send the customer to your own checkout to pay. The payment and the order stay in the system you already use.

How far that goes depends on your ordering system. With some, the link can arrive with every item already in it. With others, the assistant writes the order out for the customer to enter on your ordering page.

See how Square restaurants turn DMs into checkout orders, and how Toast restaurants take orders from Instagram DMs.

Leads: catching the job while you're on another one

For a plumber, electrician or HVAC company, the revenue moment is the callback. A customer who asks "how much to fix a leaking water heater?" wants a price and a time. An assistant can give your listed or "starting at" price, collect what the job is, where and when, and ask for a name and a phone number. You set the final price and the time when you call back.

See how contractors answer DMs and capture leads while on a job.

What the research says about AI in service

A large study of AI in customer service looked at AI helping people, not replacing them. Erik Brynjolfsson, Danielle Li and Lindsey Raymond studied 5,179 customer support agents at one company, in an NBER working paper revised in November 2023. Agents with an AI assistant resolved 14% more issues per hour on average, and novice and low skilled agents 34% more. The paper also found that AI assistance improved customer sentiment.

When AI answers customers directly, not every customer comes away helped. In Qualtrics XM Institute's study of 20,001 consumers in 14 countries in the third quarter of 2025, nearly 1 in 5 consumers who used AI for customer support saw no benefit at all.

And customers have clear worries. In Gartner's survey of 5,728 customers in December 2023, the top concern about AI in customer service was that reaching a person would get harder. AI displacing jobs came second and AI giving wrong answers third.

Our takeaway, not a finding of these studies: AI is safest when it answers from facts you gave it, and when a person is one step away.

What AI should not handle on its own

Some messages cost more when a machine answers them. Hand these to a person, every time:

• Allergies and dietary safety. A wrong answer can hurt someone. The assistant should say it can't make that call and connect the customer with your team.
• Complaints. Only a person can decide how to make it right.
• Refunds and credits. Only you decide what to give back.
• Final quotes. A listed or "starting at" price is fine. The final number is yours, after you see the job.
• Emergencies, like a gas smell or a burst pipe. The customer needs safety information and a fast call from a person.
• Anything you haven't written down. If the answer isn't in your information, the assistant should say so and pass it on, not guess.

How to tell if it's paying off

Don't take anyone's word for it, including ours. Measure it in your own business for a month.

1. Count the conversations that happen outside your open hours.
2. Count the orders or carts built in the chat, and how many were paid.
3. Count the leads captured with a name and a phone number or email, and how many became jobs.
4. Count the handoffs to a person, and read a sample to check the assistant passed on the right ones.
5. Compare what it brought in with what it costs.

Before you start, estimate what missed messages cost you now with the missed message cost calculator.

Set it up so it can't guess

The part of the risk you control is what the assistant is allowed to say. Give it your menu or services with current prices, your hours including holidays, your service area, and your policies. Tell it which topics always go to a person. Make sure it tells customers it is automated. DM DESK is built around those rules: it answers from your real information and hands allergies, complaints and anything sensitive to you.

Sources

1. Zendesk, Contextual Intelligence Becomes the New Standard for Exceptional Customer Experience in 2026 (CX Trends 2026), 2025.
2. National Bureau of Economic Research, Generative AI at Work (Working Paper 31161), Brynjolfsson, Li and Raymond, 2023.
3. Qualtrics, Businesses Risk $3 Trillion in Sales From Poor Customer Experiences as Consumers Cut Spending, 2025.
4. Gartner, Gartner Survey Finds 64% of Customers Would Prefer That Companies Didn't Use AI For Customer Service, 2024.