
AI Agents for Small Business: Real Use Cases
Six jobs a small business can hand to an AI agent, with the arithmetic shown so you can see which ones pay and which ones only look clever.
Short answer
Small businesses get real value from agents on a handful of jobs: answering common questions, drafting quotes, sending reminders, chasing invoices and replying to reviews. In our illustrative example of 27 saved hours a month, quotes and enquiries produce about two-thirds of the value. Published tool costs are small, from USD 0.99 per resolved chat on Fin.
Which jobs pay, and which only look good?
Quotes and customer questions pay best. Reminders and review replies save only minutes. The table below uses an invented service business with six staff and shows the arithmetic, so you can swap in your own volumes. Every figure in the volume and time columns is an illustrative assumption. The costs come from published prices, checked in November 2026.
| Job | Per month | Manual time | With an agent or workflow | Hours saved | Value at EUR 20 an hour |
|---|---|---|---|---|---|
| Answering common questions on chat | 200 chats, 5 minutes each | 16.7 hours | Agent resolves 60%, so 120 chats leave the queue | 10.0 | EUR 200 |
| Drafting quotes from a price list | 30 quotes, 25 minutes each | 12.5 hours | Draft in minutes, 8 minutes to check | 8.5 | EUR 170 |
| Appointment reminders and changes | 120 appointments, 3 minutes each | 6.0 hours | About half a minute of checking each | 5.0 | EUR 100 |
| Chasing overdue invoices | 25 invoices, 8 minutes each | 3.3 hours | 2 minutes to check each | 2.5 | EUR 50 |
| Replying to reviews | 15 reviews, 6 minutes each | 1.5 hours | Draft reply, 2 minutes to edit | 1.0 | EUR 20 |
| Total | 27.0 | EUR 540 |
Quotes and chat questions account for 18.5 of the 27 hours, about two-thirds. Reminders, invoices and reviews together save under 9 hours. Build the first two before the others.
What does the tool cost against that saving?
Costs are small against the value, but not trivial against the thin jobs. Using published prices:
- Chat agent. Intercom's pricing page lists its Fin agent at USD 0.99 per outcome, with helpdesk seats from USD 29 a month billed annually (USD 19 for new customers on Intercom's current offer). At 120 resolved chats that is about USD 119 a month.
- Reminders and invoice chasing. A general automation platform handles these. Zapier's Professional plan lists USD 19.99 a month billed annually for 750 tasks. The 120 reminders and 25 invoice nudges here use well under that, at about two tasks each.
- Quotes and review drafts. A business plan of an AI assistant at about USD 20 per user a month does both.
Total published cost is roughly USD 159 a month against EUR 540 of staff time, a net of about EUR 380 if you ignore the exchange rate, before setup and checking. Take away the quote and chat jobs and the saving falls to EUR 170 against about USD 40 of cost. The cheap jobs are still positive, but they are not worth a project of their own.
Job 1: answering customer questions on chat and WhatsApp
An agent answers from your own written material and hands over when it does not know. It works because most small businesses get the same ten questions. Intercom counts an outcome when a customer confirms resolution, does not ask for more help, or a workflow completes, so a customer who simply leaves is counted as resolved. Tidio's page lists a standalone Lyro agent add-on from USD 32.50 a month for 50 conversations. Check real resolution by asking for a thumbs-up after each answer.
Meta's developer documentation describes a Meta Business Agent for WhatsApp billed per token at USD 2.00 per million tokens, which it puts at roughly 4 to 5 cents a message. See WhatsApp automation for business for the policy rules that apply.
Job 2: drafting quotes
An assistant fills your quote template from a price list and flags anything unusual, and a person approves the price. It is the best return on this page because quotes are slow and the draft can be checked line by line. Keep the price list in one document, tell the assistant to use nothing else and make a person sign off every quote. A wrong quote is a promise.
Job 3: answering the phone
A voice agent answers calls, asks a few questions and books or takes a message. Goodcall's published plans start at USD 79 a month for 100 unique customers, about 150 calls. Its value comes from captured jobs rather than staff hours, so we leave it out of the table. Our AI receptionist guide works through the missed-call arithmetic.
Jobs 4 to 6: reminders, invoices and reviews
These are worth doing only because they are cheap. Send appointment reminders by rule, with no AI at all. Chase overdue invoices with two or three scheduled, polite messages, and let a person phone any long-standing customer. Draft review replies with an assistant, edit them, and post them yourself. The honest benefit of invoice chasing is cash arriving sooner, not hours saved, and we have not put a figure on that.
Where a person must stay in charge
Keep a person in charge of prices, refunds, complaints, health or legal questions, and the first reply to a negative review. An agent can prepare each of these. It should not send them. Give it the least access it needs, such as draft rights instead of send rights, and log what it does.
Which rules apply to a small business running agents?
Three duties apply to almost every customer-facing agent. This is general information, not legal advice.
- Tell people it is an AI. Article 50 of the EU AI Act has applied since 2 August 2026. It requires chatbot providers to design systems so people are informed they are interacting with an AI, unless that is obvious. Put it in the first message.
- Sign a processing contract. Under GDPR Article 28, any vendor that handles customer personal data for you needs a contract that limits what it can do with the data. Ask for the vendor's data processing agreement before you connect anything.
- Always offer a person. A chat or phone agent should give a clear route to a human, in the first message and again whenever it cannot help.
How to start
Pick the quote job or the chat job, run it for a month with approval, and count real minutes. Our guide to AI agents for business explains when an agent beats a plain workflow, and the AI agents examples page lists more by department. If you want a second opinion on which job to start with, see our AI automation service. We are a new studio and we quote in writing.
Sources
Frequently asked questions
How can I use AI for my small business?
Start with the jobs that repeat daily and have a checkable answer: replying to common questions, drafting quotes from a price list, sending reminders and chasing late invoices. Measure the minutes before and after, and keep a person approving what customers see.
How do I use AI for small business marketing?
Use it to draft review replies, summarize customer feedback and prepare first versions of posts and emails. Edit everything before it goes out. Marketing claims about your own business need your fact-checking, because a model will happily invent details.
What is an AI voice agent for a small business?
It is software that answers your phone, asks the caller questions, books or takes a message and passes urgent calls to a person. Goodcall, for example, lists plans from USD 79 a month for 100 unique customers.
Do small businesses really need AI agents?
Not usually, and not first. Plain automation handles reminders and data entry more cheaply and predictably. Agents earn their cost on jobs that need reading and writing free text, such as enquiries and quotes, at a volume that makes the time add up.
Does Meta offer an AI agent for small businesses?
Meta's developer documentation lists a Meta Business Agent for WhatsApp, billed per token at USD 2.00 per million tokens, which it puts at roughly 4 to 5 cents a message. Check the current terms before relying on it.
Keep reading
AI Agents for Business: What They Can Do Today
What AI agents can reliably do for a business today, what they get wrong, how vendors charge per conversation, and a test for agent or workflow.
6 min read
AI Agents Examples for Business
Eight AI agent examples for business by department, with what each decides, what access it needs, where a person checks and the published price.
5 min read
AI Receptionist for Small Business: Cost, Setup and Limits
When an AI receptionist pays for itself, what setup involves, where it fails, and the GDPR and recording rules to check first. Published prices included.
6 min read
Free auditLet us tell you what to fix first.
Send us your website and we will send back a short, honest audit of what is holding it back.
Request a free audit