Is a chatbot worth it for small business? A break-even guide

Whether a chatbot is worth it for a small business comes down to one sum: the time it saves you plus the leads it catches, against what it costs each month. If your site gets a steady flow of questions that your own pages could answer, it usually pays for itself. If your site gets a few visitors a week, or every question needs someone to look up an account, it usually doesn't. You can work out which side you're on in about ten minutes with numbers you already have.

We sell a chatbot, so take this as a vendor's guide that tries to be fair. Below is the break-even sum, two worked examples with made-up numbers, the businesses we'd tell not to bother, and a way to test it for three days without paying.

The break-even sum

You need four numbers. Estimate them from your own inbox, phone and calendar, not from anyone's averages.

1. Questions per week. Count the emails, contact-form messages, DMs and calls in a normal week that a well-written page could have answered. Shipping times, opening hours, service area, prices, "do you do X?". Leave out anything that needs you to check an order or a diary.
2. Minutes per answer. How long it takes you to read, reply and get back to what you were doing. Be honest: a two-line email reply often costs five minutes once you count the interruption.
3. The value of a lead. What a new customer is typically worth to you in profit on the first job or order, multiplied by the share of enquiries that turn into customers.
4. The plan price. Ours start at $49 a month. See our pricing for the rest.

Then:

• Time value per month = questions per week × 4.3 × share the bot can answer × minutes per answer ÷ 60 × what an hour of your time is worth.
• Lead value per month = extra leads per month × value of a lead.
• Worth it if time value + lead value is more than the plan price.

"Share the bot can answer" matters. A chatbot answers from what's written on your site, so if half your questions are about things your pages cover and half need you personally, use 50%. Don't assume it takes every question off your plate.

"Extra leads" also matters. A lead who would have filled in your contact form anyway isn't extra. It's the visitor who had one question at 10pm and would otherwise have left. We cover how to measure that difference in chatbot vs contact form.

Example 1: a shop with a steady flow of questions

These numbers are invented to show the sum. They aren't data from any customer.

Say an online shop gets 30 product and delivery questions a week. About half are things already on its pages: shipping times, returns, sizing. The owner spends about 5 minutes on each and values their time at $25 an hour.

• Questions the bot could take per month: 30 × 4.3 × 50% ≈ 65
• Time saved: 65 × 5 minutes ≈ 325 minutes, or about 5.4 hours
• Time value: 5.4 × $25 ≈ $135 a month

That's more than our $49 Starter plan before counting a single lead. Volume isn't a problem either: one question can take two or three replies once the visitor follows up, but 65 questions is still a few hundred replies, well inside Starter's 1,000.

For this shop the more useful question is which chatbot, not whether. Our AI chatbot cost guide prices the same volumes across vendors.

Example 2: a quiet service business

Again, invented numbers for illustration only.

Say a one-person landscaping business gets 4 website questions a week. Half are about service area and whether they do small jobs. Five minutes each, $25 an hour.

• Questions the bot could take per month: 4 × 4.3 × 50% ≈ 9
• Time saved: 9 × 5 minutes = 45 minutes
• Time value: 0.75 × $25 ≈ $19 a month

On time alone, that falls $30 short of $49. So the decision rests on leads. Say a typical first job brings $400 profit and one in four enquiries becomes a customer. Each lead is then worth about $100.

• Leads needed to break even: $49 ÷ $100 ≈ 0.5 a month

One extra lead every two months covers the cost, once the time saved is counted too. Whether a chatbot actually produces that extra lead on a site this quiet is the part you can't know in advance. That's what the trial is for. If after a fair test it isn't catching anyone your contact form wouldn't have, it isn't worth it for you, and that's a perfectly good answer.

What lowers the bar

On a monthly plan, our first three months are 30% off, so Starter is $34.30 a month while you find out. Paying yearly gets 12 months for the price of 10, which works out to about $40.83 a month on Starter. Neither changes the sum much, but both are worth knowing if you're near the line.

If price per reply is what decides it, we aren't the cheapest. As of October 2026, Botsonic sells 1,000 messages for $19, SiteGPT sells 4,000 messages on its cheaper model for $59, and Chatling sells 1,750 credits for $25. All three cost less per message than our Starter. Each charges extra to remove its logo, which we don't, and our comparisons with Botsonic, SiteGPT and Chatling set out that trade in full. To cost your own volume across vendors, use the chatbot cost calculator.

Who a chatbot is worth it for

The businesses where the sum tends to come out well share a few traits:

• The same questions, over and over. Delivery, returns, hours, service area, "do you take my insurance?", "do you do weekend work?". If you've ever thought "it says that on the website", a chatbot can answer it.
• The answers are written down. The bot answers from your pages. If your policies live only in your head, write them down first. That alone often cuts the questions.
• Visitors arrive when you're busy or closed. Evenings, weekends, or while you're on a job. A chatbot answers at those times and offers to take a name and email for the questions it can't answer. That's what lead capture does.
• A question stands between the visitor and the sale. "Is this in stock in medium?" or "Do you work in my town?". If the answer is on your site but hard to find, the bot surfaces it.

Restaurants, salons, contractors and online shops often fit this pattern. Our pages for contractors, restaurants, salons and ecommerce list the questions each one gets most.

Who it isn't worth it for

We'd rather you didn't sign up than cancel in a month. A chatbot, ours included, is probably not worth it if:

• Your site gets very little traffic. If a handful of people visit a week and almost nobody asks anything, there's nothing for the bot to do. A clear FAQ page and a contact form cost nothing and will do the job.
• Your questions are all about the customer's own account. "Where's my order?", "When's my appointment?", "What's my balance?". Our chatbot can't see orders or customer accounts, and it doesn't book appointments. If most of your questions are like that, you need a tool connected to your order or booking system, not a website chatbot.
• You want to chat with customers yourself. We don't offer live chat with humans: there's no inbox for you or your staff to take over a conversation. If that's what you want, tools built around a live chat inbox, such as Tidio, Crisp or Intercom, suit you better.
• The questions need a professional's judgement. Our chatbot doesn't give legal advice, and it isn't a HIPAA-covered system: we don't sign Business Associate Agreements. Law firms and dental practices can still use it for general front-desk questions, but not for case details or health information. We cover where that line sits in our posts on law firm chatbot ethics and dental chatbots and HIPAA.
• Your customers would rather not talk to a bot. In a Gartner survey of 5,728 customers, published July 9, 2024, 64% said they would prefer companies didn't use AI in customer service. If your customers are mostly people who want to call you, keep your phone number front and centre and think hard before adding a chat window.

How to test it cheaply with the 3-day trial

Our free trial gives you 3 days of every Growth feature and 100 replies, with no card required. That's short on purpose: long enough to put the bot on your real site and see what real visitors ask, with a date by which to decide. Here's how to get a usable answer out of it.

Before you start the clock

1. Write down your four numbers from the break-even sum.
2. Check your site actually answers your common questions. If your returns policy, service area or hours aren't on a page, add them.
3. Pick a time to start when you'd normally get questions, not a public holiday.

During the three days

1. Sign up, give the bot your site, and add the widget. The install steps for WordPress, Shopify, Squarespace and Wix each take a few minutes.
2. Ask it ten of your real customer questions yourself, including a couple it shouldn't be able to answer. It should say it doesn't know and offer to take details, not guess. Our post on stopping a chatbot making up policies has a fuller test.
3. Leave it running and don't touch it.

At the end, look at three things

1. Replies used. Multiply by 10 for a rough month. That tells you which plan you'd need, or that you'd barely use one.
2. What people asked. The answer gaps list shows every question it couldn't answer. If those are things you could add to your site, the bot gets more useful. If they're all "where's my order?", see the section above.
3. Leads caught. Did anyone leave details, and would they have contacted you anyway?

Put those back into the sum. Three days won't prove anything with certainty, especially on a quiet site, but it replaces guesses with your own visitors' behaviour. If it doesn't add up, let the trial end. Nothing is charged.

Frequently asked questions

Is a chatbot worth it for a small business?

It is when your site gets regular questions that your pages could answer, and the time saved plus the extra leads come to more than the monthly price. It usually isn't for very quiet sites, for businesses whose questions all need an account looked up, or for owners who want to chat with customers themselves. Work it out with your own numbers rather than anyone's averages.

How many questions does my site need for a chatbot to pay off?

There's no fixed number, because it depends on how long each answer takes you and what a lead is worth. As a rough guide, if your site gets only a few questions a month and you rarely lose a customer for want of a quick answer, a chatbot is unlikely to earn its cost. If you answer the same questions several times a day, it very likely will.

Will a chatbot replace my contact form or phone line?

It shouldn't. A chatbot is best at answering common questions and catching visitors who'd otherwise leave. Keep your contact form, phone number and email visible, because some customers prefer them and some questions need a person.

Can I try a chatbot without paying?

Yes. Our free trial runs for 3 days with 100 replies and every Growth feature, and needs no card. Several other vendors offer free plans, usually with their logo shown and a small allowance. Whichever you try, put it on your real site so you see real questions.

What does a small business chatbot cost?

For hosted website chatbots, roughly $19 to a few hundred dollars a month as of October 2026, depending on volume, lead capture and whether you pay to remove the vendor's logo. Our plans are $49, $149 and $449 a month. Our cost guide works through full bills across vendors.