A small business chatbot costs $0 to $100 a month on a no-code platform, $100 to $500 a month on a support platform, and $10,000 to $150,000 custom-built. The real 2026 price bands, and the fees hiding behind them.

How Much Does an AI Chatbot Cost for a Small Business?

An AI chatbot costs $0 a month. It also costs $150,000. Both numbers are real, and neither one answers the question a small business owner is actually asking, which is: what will it cost me, for my website, with my volume, to get a chatbot that answers customers properly and captures leads instead of annoying them?

The answer depends on which of three things you buy. A no-code platform subscription, a full support platform with AI built in, or a custom build. Those are three different products with three different price shapes, and most of the confusion in chatbot pricing comes from comparing them as if they were the same thing.

This guide gives you the 2026 price bands from published pricing guides and platform price checks, explains the pricing models that make a cheap plan expensive, covers what a build actually takes in time, and shows how to work out whether a chatbot pays for itself. All prices are in US dollars unless a source is labelled otherwise, because chatbot platforms bill in USD even when your customers are in Canada.

The short answer

For a small business in 2026, an AI chatbot usually costs one of the following, depending on the route:

• No-code platform (Tidio, ManyChat, Chatbase, Chatling): $0 to $50 a month to start, with most small businesses landing at $25 to $100 a month once AI features and real conversation volume are counted. Setup takes days, not weeks, and no developer is needed.
• Full support platform with AI (Intercom, Freshchat, Crisp): $100 to $500 a month for a small team, once seats, AI usage, and add-ons are counted. The advertised entry price is usually $19 to $45 a month, and the real bill is two to four times that.
• Custom-built chatbot: $10,000 to $30,000 for a basic build, $30,000 to $150,000 for an agency-built system with integrations, and $75,000 to $150,000 or more for an advanced AI build. Annual maintenance runs 15 to 20 percent of the build cost, plus $100 to $500 a month in AI model fees.

Those bands come from price checks done in 2026. Bluehost's 2026 pricing guide puts a small business chatbot at $20 to $100 a month, a basic custom build at $10,000 to $30,000, and an advanced AI build at $75,000 to $150,000 and up. Elfsight's 2026 guide puts the entry tier at $0 to $50 a month, the mid-tier at $100 to $500 a month, and agency-built chatbots at $30,000 to $150,000, with enterprise systems reaching $200,000 to $1,000,000. FwdSlash's comparison, with every figure verified at the vendor's pricing page in September 2026, puts the budget band at $17 to $29 a month and the mid-range at $30 to $50 a month for small business tools.

The shape is consistent across every source: the subscription is cheap, the build is expensive, and the gap between the advertised price and the real monthly bill is where most small businesses get surprised. The rest of this guide is about that gap.

What actually sets the price

Before comparing platforms, it helps to know the four things that decide what you pay. Two chatbots can look identical in a demo and cost ten times apart, because they differ on these four.

How the chatbot is priced. There are three pricing models, and they behave very differently as you grow. Flat-rate plans charge one monthly fee for a set allowance, which makes the bill predictable. Per-seat plans charge per team member, so the bill grows every time you hire. Per-conversation or per-resolution plans charge for each chat the AI handles, so the bill grows with your traffic and spikes in your busy season. Intercom's Fin, for example, charges $0.99 per resolved conversation on top of seat fees, which is cheap at 50 conversations a month and $495 a month at 500 resolutions. The model matters more than the starting price, because it decides what your bill looks like in six months.

How many conversations it handles. Almost every platform prices by volume: conversations, messages, credits, or contacts per month. A free plan that covers 50 conversations a month is generous for a business getting 40 chats a month and useless for one getting 400. The allowance is the real price. A $25 plan with 100 conversations costs more per conversation than a $40 plan with 700, once you pass the limit and start paying overage or upgrading tiers.

What it connects to. A chatbot that only answers questions from a FAQ page is cheap. A chatbot that checks order status, books appointments, qualifies leads into your CRM, and hands off to a human with full context is a different product. Each integration, a calendar, a CRM, a helpdesk, a payment system, adds setup time and usually a higher plan tier, because integrations sit behind the mid-range and premium plans on most platforms.

Who builds it. A platform you configure yourself costs your time. A platform with managed setup costs a setup fee or a higher tier. A custom build costs a developer or an agency, and that is where the five-figure numbers come from. The build cost is a one-time payment, but the maintenance is not, and the maintenance is the number most quotes leave out.

Route 1: No-code platforms, $0 to $100 a month

This is where most small businesses should start. You pick a platform, connect it to your website, train it on your website pages and FAQ, and it answers visitors. No code, no developer, live within days.

The entry prices, verified on vendor pricing pages in September 2026, look like this:

• Tidio: free plan with 50 conversations a month. Starter from $24.17 a month on annual billing, covering 100 conversations. The Lyro AI add-on runs about $32.50 to $39 a month on top of the base plan.
• ManyChat: free plan for basic automation. Essential at $17 a month, and Pro at $39 a month, which is the first tier with AI features. A GPT-powered add-on costs a further $29 a month on some plans.
• Chatbase: free plan with 50 message credits. Hobby at $40 a month for 700 message credits, with Standard at $150 and Pro at $500 for heavier use.
• Chatling: free plan with 200 AI credits and unlimited chats. Starter at $25 a month, Growth at $75, and Scale at $295.
• FwdSlash: free plan with 200 messages a month. Intermediate at $20 a month for 2,000 messages, which is the cheapest AI-included price in the September 2026 comparison.
• Tawk.to: unlimited live chat agents free, forever. The AI add-on costs $29 a month.
• ChatBot.com: from $19 per user a month on annual billing, according to Bluehost's 2026 guide.
• Botpress: free plan with 100 conversations a month, then from $150 a month, which puts it at the top of this route and the bottom of the next one.

A realistic small business budget on this route is $25 to $100 a month. The free plan works while volume is low, and the first paid tier covers most businesses that get a few hundred chats a month. The trap is the add-on stack, covered in the hidden costs section below: a $29 base plan plus a $39 AI add-on plus a $29 automation add-on plus $20 to remove branding is $97 a month, not $29.

This route fits a business that wants a chatbot on its website within a week, with lead capture and FAQ answers, without hiring anyone. It fits poorly when the chatbot has to work across WhatsApp, Instagram, email, and a helpdesk at the same time, or when it has to take actions inside your internal systems, because those are the jobs that push you to the next route.

Route 2: Full support platforms with AI, $100 to $500 a month

These are customer service platforms first and chatbots second. Intercom, Freshchat, Crisp, and Zendesk give you a shared inbox, a help centre, ticketing, and live chat, with an AI agent sitting on top. You pay for the platform and the AI separately, which is why the bill is harder to predict.

Intercom and Fin. Intercom's seat plans in 2026 are Essential at $29 per seat a month on annual billing ($39 monthly), Advanced at $85 ($99 monthly), and Expert at $132 ($139 monthly). Fin, the AI agent, is billed separately at $0.99 per outcome, charged at most once per conversation, according to Gleap's pricing check from September 2026 and Voiceflow's Intercom guide. On a non-Intercom helpdesk there is a 50-outcome monthly minimum, about $49.50, and qualifying a lead costs $9.99 per qualification. A three-person team on Essential with 300 Fin resolutions a month pays about $87 in seats plus $297 in resolutions, or roughly $384 a month, for a product advertised at $29 a seat.

Freshchat. Growth is $19 per agent a month on annual billing, with 500 Freddy AI sessions included. It is the cheapest entry point on this route, and the AI allowance is the thing to watch: once the included sessions run out, the meter starts.

Crisp. Mini is $45 a month per workspace with 4 seats, Essentials is $95 a month with 10 seats and a knowledge base, and Plus is $295 a month. Crisp is the only platform on this route with genuinely flat pricing, which is why it suits small teams that want a predictable bill.

The realistic band. Elfsight's guide puts mid-tier SaaS platforms at $100 to $500 a month, and a team of 3 to 5 people handling 2,000 to 5,000 monthly conversations at a realistic $150 to $800 a month once features and usage are counted. That is the honest number for this route: the advertised entry price is a floor, not a bill.

This route fits a business that already has a support team, handles real ticket volume, and wants the chatbot, the inbox, and the help centre in one place. It fits poorly when the chatbot is the only thing you need, because you pay for a platform you will not use, and when your volume is seasonal, because per-resolution pricing makes a busy month cost three times a quiet one.

Route 3: Custom-built chatbots, $10,000 to $150,000 and up

A custom build means a chatbot designed around your business: trained on your documents, connected to your CRM and calendar, with conversation flows your team designed, and hosted on your own stack. Nobody else's customers share it, and nobody else's roadmap decides what it can do.

The cost bands, from 2026 guides:

• Basic custom chatbot: $10,000 to $30,000, according to Bluehost. A focused FAQ and lead-capture bot, one channel, no deep integrations.
• Mid-range custom build: $30,000 to $150,000 for an agency-built system, according to Elfsight. Layer3Labs puts an LLM-powered bot with CRM integration at $10,000 to $35,000, and the full custom range at $2,000 to $100,000 and up, which shows how wide the band is: the scope decides the price, not the label.
• Advanced AI build: $75,000 to $150,000 and up, according to Bluehost, for a system with retrieval over large document sets, several integrations, and custom workflows. Enterprise deployments reach $200,000 to $1,000,000.

Maintenance is the number quotes hide. Annual maintenance runs 15 to 20 percent of the initial build cost, so a $50,000 chatbot needs $7,500 to $10,000 a year just to stay current. On top of that, Layer3Labs puts LLM API fees at $100 to $500 a month for a typical small business build. A $30,000 build therefore costs roughly $4,500 to $6,000 a year in maintenance plus $1,200 to $6,000 a year in model fees, every year, for as long as it runs. Budget the second year before you approve the first.

What the money buys. A custom build buys control: your data stays in your systems, the chatbot speaks in your voice because it was trained on your material, and it can take actions, check an order, book a slot, update a record, instead of only answering questions. It also buys a slower start and a real project: discovery, knowledge base preparation, conversation design, integrations, testing, and launch, with your team involved at every stage.

This route fits a business with volume or complexity that platforms cannot handle, a regulated industry where data cannot leave your systems, or a chatbot that is central to how the business sells. It fits poorly when the job is answering the same 30 questions and capturing leads, because a $40-a-month platform does that job at 1 percent of the price.

The hidden costs that double the advertised price

The advertised price is a starting point, not a bill. These are the five charges that quietly push a $29 plan toward $100, and a $100 plan toward $300.

AI as an add-on. On several platforms the base plan is a live chat tool and the AI is a paid extra. Tidio's Lyro AI, ManyChat's GPT add-on, and Tawk.to's AI add-on all sit on top of the base price. Check whether the plan you are looking at includes AI answers, or only the chat window the AI lives in.

Per-conversation billing. Some platforms charge per AI conversation rather than a flat monthly fee. At low volume this is cheap. At a few hundred conversations a month it adds up fast, and at seasonal peaks it can triple the bill without warning. Per-resolution pricing has the same shape: Intercom's Fin at $0.99 per outcome costs $49.50 for 50 resolutions and $495 for 500.

Branding removal. Free and starter plans usually show the platform's branding on your chatbot. Removing it costs extra, about $20 a month on Tidio's add-on stack. If your chatbot is a customer-facing surface, which it is, budget for this from the start.

Add-on stacking. Tidio is the clearest example: Starter at $29 a month becomes $97 a month once you add Lyro AI ($39), Flows automation ($29), and branding removal ($20). Each add-on is reasonable on its own. Together they triple the bill. Price the stack you actually need, not the tier you start on.

Setup and training. Managed setup, where the vendor configures the platform for you, is sometimes free and sometimes a separate fee. The real cost is usually your time: writing the knowledge base, answering the same questions the chatbot will answer, and testing it against real customer questions. A chatbot trained on a thin FAQ gives thin answers, and no platform fixes that.

How long does it take to build a chatbot?

The timeline depends on the route, and the differences are large enough to matter commercially. If a competitor already has a chatbot answering customers while you are still in week six of a build, the timeline is a cost, not a detail.

The sources agree on the shape. Hyperleap's build-versus-buy guide puts self-serve platforms at 1 to 7 days, managed setup at 3 to 5 days, and custom agency builds at 8 to 16 weeks. AutomateX360's timeline puts a focused bot at about a week, a lead qualification bot at 1 to 3 weeks, and a custom LLM build at 4 to 8 weeks or more. Layer3Labs puts a production custom release at 4 to 12 weeks, with discovery in weeks 1 to 2, the core build in weeks 3 to 6, integrations in weeks 7 to 9, and hardening and rollout in weeks 10 to 12. Ongito's deployment guide puts a WhatsApp chatbot for a small business at 2 to 4 weeks, or 6 to 8 weeks when the account is new, a CRM integration is involved, and the content has to be written from scratch.

The most common delay is not engineering. It is content: the business has not written down the answers the chatbot needs to give. If your FAQ, pricing, policies, and service descriptions already exist in one place, a platform chatbot can go live in days. If they live in three people's heads, add two weeks while someone writes them down.

Build or buy: which route fits your business

The choice is simpler than the market makes it look. It comes down to two questions: how much volume do you have, and how specific is the job?

Buy a platform when: you want a chatbot on your website this month, your questions repeat, your volume is under a few thousand conversations a month, and your team can configure a tool without a developer. This covers most small businesses. Start on a free plan or the cheapest AI-included tier, and upgrade when the volume proves the need.

Buy a support platform when: you already have a support team, you handle tickets across email and chat, and you want the chatbot, the inbox, and the help centre in one system. The per-seat and per-resolution pricing hurts at low volume and makes sense at high volume, because the AI resolves tickets that would otherwise need a person.

Build custom when: the chatbot has to take actions inside your systems, your data cannot leave your infrastructure, your conversation volume is high enough that per-resolution pricing costs more than a build within two years, or the chatbot is a core part of how you sell. Run the maths first: a $50,000 build plus $10,000 a year in maintenance breaks even against a $500-a-month platform in about eight years, and against a $100-a-month platform it never does.

Do neither yet when: you do not know what your customers ask most, because the first job is to read a month of emails and chats and write the answers down. That list is the knowledge base, and it is free to make. A chatbot built on a guessed FAQ is a chatbot that guesses at customers.

Our guide to what AI automation agencies do covers the wider picture: a chatbot is one automation among several, and the agencies that build them well start with the business problem, not the tool. Our comparison of AI agents and chatbots explains the difference that decides which one you actually need: a chatbot answers, an agent acts, and they are priced differently for that reason.

Does a chatbot pay for itself?

A chatbot is not a website feature. It is a member of staff that works nights and weekends, and the only question that matters is whether it earns more than it costs. Three numbers decide that.

How many conversations it resolves without a human. Intercom reports an average resolution rate of 67 percent across Fin customers, with top deployments reaching 80 percent or more, but those figures are vendor-reported rather than independently audited. LiveAgent's comparison notes that independently reported customer case studies put real-world Fin resolution rates closer to 42 to 50 percent. The honest expectation for a well-trained small business chatbot is that it fully resolves roughly half of routine questions, and hands the rest to a person with context. That is still a large saving if your team currently answers those questions one by one.

How fast it answers. A visitor who waits is a visitor who leaves. Our guide to the five-minute rule covers why responding within five minutes changes conversion: a chatbot answers in seconds, at any hour, which is the one thing no human team can match without night shifts. For a business that gets enquiries outside working hours, that alone can justify a $40-a-month plan.

What a resolved conversation is worth. A chatbot that books a consultation for a service business, or recovers an abandoned cart for a store, pays for itself in a handful of conversions. A chatbot that only answers opening-hours questions saves staff time, which is real but harder to count. Before you buy, write down the one job the chatbot has to do, and the value of one completed job. If ten booked calls a month are worth $2,000 to you, a $100-a-month chatbot is cheap. If the chatbot's job is answering questions that never led to a sale, it is expensive at any price.

A 2025 Salesforce study, cited in Bluehost's pricing guide, found that AI handled around 30 percent of customer service cases in 2025, and service professionals expect that figure to reach 50 percent by 2027. The direction is clear. The question for a small business is not whether chatbots work, but whether the specific chatbot you buy does the specific job you need, at a price the job justifies.

What to check before you sign

Most chatbot regret comes from a mismatch between the plan and the business, not from the technology. Work through this list before you pay for anything.

Start on the free plan, and measure real volume first. Every platform above offers a free tier or a trial. Run it for two to four weeks and count actual conversations. The plan you need is decided by your real numbers, not by the tier the sales page recommends.

Price the stack, not the tier. Add the AI add-on, the branding removal, the automation add-on, and the next volume tier together. If the real monthly cost is more than double the advertised price, say so in your budget, and compare platforms on the real number.

Ask how overage works. What happens when you pass the conversation limit: does the chatbot stop answering, does it charge per extra conversation, or does it force an upgrade to the next tier? A platform that stops answering at your busiest moment is worse than a more expensive one that keeps going.

Ask who owns the training data and the conversation history. If you leave the platform, can you export the knowledge base, the conversation logs, and the lead list? A chatbot that holds your customer conversations hostage is a lock-in problem, not a tool.

Check the human handoff. The chatbot must hand a conversation to a person cleanly, with the full transcript, and the person must be able to take over from a phone. Test this before launch, not after a customer complains that they were stuck in a loop.

Read the data and privacy terms. A chatbot that sends your customer conversations to a third-party AI model has privacy implications, and in some industries it has legal ones. Know where the data goes before the first customer types a question.

Confirm the cancellation terms. Month-to-month or annual? If annual, what happens if the chatbot does not work out in month three? The cheapest annual price is expensive if you are locked into a tool you stop using.

Questions to ask before hiring a chatbot agency or developer

If you go the custom route, take this list to every call. The quality of the answers matters more than the smoothness of the pitch:

1. What exactly is included in the build price, and what is billed separately?
2. How many integrations are included, and what does each extra integration cost?
3. Who writes the knowledge base, you or us, and how many hours of content work is included?
4. What is the realistic launch timeline for our scope, and what do you need from us to hit it?
5. What does annual maintenance cover, and what counts as a chargeable change?
6. What are the ongoing AI model fees, and who pays them?
7. What resolution rate do your existing clients see, and can we talk to one of them?
8. What happens if the chatbot gives a wrong answer: who is liable, and how is it corrected?
9. Who owns the chatbot, the training data, and the conversation history when the contract ends?
10. What does the handoff to a human look like, and can our team take over from a phone?

A builder who answers all ten clearly is rarely the cheapest quote, and is almost always the cheapest chatbot once the missing pieces are counted.

Frequently asked questions

How much does an AI chatbot cost for a small business?

In 2026, a small business can start with an AI chatbot for $0 to $50 a month on a no-code platform, with most serious setups landing at $25 to $100 a month once AI features and real conversation volume are counted. A full support platform with AI runs $100 to $500 a month for a small team. A custom-built chatbot starts around $10,000, with agency builds commonly at $30,000 to $150,000. Annual maintenance adds 15 to 20 percent of the build cost.

How much does a chatbot cost per month?

For a small business, the realistic monthly cost is $25 to $100 for a no-code platform chatbot, $100 to $500 for a support platform with AI, and $0 on a free plan while volume stays low. The advertised entry price is usually lower than the real bill, because AI add-ons, per-conversation charges, and branding removal sit on top of the base price. Price the full stack before you budget.

How much does it cost to build a chatbot?

A self-serve platform chatbot costs $0 to $200 a month and goes live in 1 to 7 days. A custom-built chatbot costs $10,000 to $30,000 for a basic build, $30,000 to $150,000 for an agency-built system, and $75,000 to $150,000 or more for an advanced AI build. Timelines run 4 to 12 weeks for most custom builds, and 8 to 16 weeks for a full agency scope.

How long does it take to build a chatbot?

A platform chatbot goes live in 1 to 7 days if your content is ready. A focused FAQ bot takes about a week. A lead qualification and booking bot takes 1 to 3 weeks. A custom LLM build with retrieval over your documents takes 4 to 8 weeks or more. A full agency custom build takes 8 to 16 weeks. The most common delay is content: writing the answers the chatbot needs to give.

How long does it take to build an AI chatbot?

The same ranges apply: days for a platform, weeks for a custom build. The AI part is rarely the slow part. The slow parts are preparing the knowledge base, connecting the chatbot to your CRM and calendar, and testing it against real customer questions. A business with its FAQ and policies already written down can launch a platform chatbot in days.

What is the cheapest AI chatbot for a small business?

The cheapest working options in 2026 are Tawk.to, which offers unlimited live chat free with an AI add-on at $29 a month, and FwdSlash, whose Intermediate plan is $20 a month for 2,000 messages with AI included. Tidio and ManyChat both offer free plans, and Chatbase offers a free plan with 50 message credits. The cheapest plan is only the right choice if its conversation allowance covers your real volume.

Is Intercom worth it for a small business?

Intercom makes sense when you already have a support team and real ticket volume, because you get the chatbot, the inbox, and the help centre in one system. At $29 a seat a month plus $0.99 per Fin outcome, a small team can easily reach $300 to $500 a month. For a business that only needs a website chatbot, a no-code platform at $25 to $100 a month does the same job for a fraction of the price.

What is the difference between a chatbot and an AI agent?

A chatbot answers questions from a script or a knowledge base. An AI agent takes actions: it checks an order, books an appointment, updates a record, and hands off to a human with context. Agents cost more to build and run, because they need integrations and guardrails, but they do jobs a chatbot cannot. Our comparison of AI agents and chatbots explains which one fits which job, and our guide to what AI automation agencies do covers the wider automation picture.

Can a chatbot replace my customer service team?

No, and a vendor who says otherwise is selling, not advising. A well-trained chatbot fully resolves roughly half of routine questions, and hands the rest to a person. What it replaces is the repetitive first half of support: answering the same questions, collecting the same details, and booking the same appointments. Your team keeps the conversations that need judgement, empathy, or a decision.

How much does a website chatbot cost?

A website chatbot is the same product as a small business chatbot, so the same bands apply: $0 to $50 a month to start on a no-code platform, $25 to $100 a month for a serious setup, and $100 to $500 a month on a full support platform. The website is usually the first channel a chatbot goes live on, and the cheapest place to prove it works before adding WhatsApp or Instagram.

What does a chatbot cost to maintain?

On a platform, maintenance is included in the subscription: you pay the monthly fee and the vendor keeps the system running. On a custom build, annual maintenance runs 15 to 20 percent of the initial build cost, plus $100 to $500 a month in AI model fees. A $50,000 custom chatbot therefore costs roughly $7,500 to $10,000 a year in maintenance, plus $1,200 to $6,000 a year in model fees, for as long as it runs.

The bottom line

An AI chatbot for a small business does not have one price in 2026, because it is not one product. It has a price per route, a price per volume, and a price per job, and only the last of those decides whether it was worth it. Start with the job: the one thing the chatbot has to do, and the value of one completed job. Then pick the cheapest route that does that job properly. For most small businesses that is a no-code platform at $25 to $100 a month, live within a week, measured on real conversations for a month before any upgrade. A support platform earns its $100 to $500 a month when you have a team and ticket volume to match. A custom build earns its $10,000 to $150,000 when the chatbot has to act inside your systems, or when per-resolution pricing would cost more than the build within two years. And remember the arithmetic: a $29 plan that needs $39 of AI, $29 of automation, and $20 of branding removal is a $97 plan. Price the stack, count your real conversations, and let the job decide the budget, not the sales page.

Our team builds websites, brands, and the AI automations around them for service businesses in Vancouver, Canada and internationally, starting with the job the automation has to do, because a chatbot is only valuable once it answers real customers, captures real leads, and hands off cleanly to a real person. If you are comparing chatbot options, start a project or book a free call and bring the quotes with you. We will help you compare them on the same job, the same volume, and the same real monthly cost, before you sign any of them.