Browse our curated list of AI Chatbot Development Companies
AI chatbot development cost in 2026 runs from about $250 for a simple FAQ bot on a website to $50,000 or more for a complex assistant. That top tier works inside your CRM, order system or helpdesk. Our AI chatbot listing shows 12 companies, but only 6 name AI chatbot development as a service on their profiles. Their average project sizes run from $250 to $25,000, with a median of $1,750.
The build is only part of the bill. A chatbot also pays for every conversation it handles, through the language model behind it. For the same traffic, that monthly bill can be $27 or $1,000. It depends on the model you pick. This guide covers both. It shows what each type of chatbot costs to build and run. It also compares US and Indian rates and explains how to read a quote.
Key takeaways
- A rule-based or template chatbot costs $250 to $2,500. A custom AI chatbot that answers from your documents costs $10,000 to $25,000. One that takes actions in your systems costs $25,000 to $50,000 or more.
- Running costs depend mostly on the model. For 10,000 conversations a month, our worked example costs about $27 on Claude Haiku 5.5 and about $540 on Claude Sonnet 5.5.
- US federal contract rates put AI engineers at a median of $175 an hour. So a mid-sized custom chatbot built at US rates comes to about $73,000.
- Of 12 companies on our AI chatbot listing, only 6 name AI chatbot development as a service. Four of the 6 average $2,500 or less per project, and 2 average $10,000 or more.
How Much Does AI Chatbot Development Cost?
Three questions set the price more than anything else. What does the chatbot need to know? What does it need to do? Where will it work? A bot that answers 30 fixed questions on one website is a small job. Now picture a bot that reads your help centre, checks an order in your ERP, books a call and replies on WhatsApp. That is different. That is a software project with several integrations to build and test.
| Chatbot type | What it does | Typical build cost | Rough timeline |
|---|---|---|---|
| Rule-based or template chatbot | Answers set questions, collects leads, routes to a person | $250 to $2,500 | 1 to 3 weeks |
| Custom AI chatbot on your documents (RAG) | Answers from your help centre, policies and PDFs, with handoff to a person | $10,000 to $25,000 | 6 to 10 weeks |
| AI chatbot that takes actions | Checks orders, books appointments, updates the CRM, works on web and WhatsApp | $25,000 to $50,000 | 10 to 16 weeks |
| Enterprise AI assistant | Single sign-on, several languages, audit logs, data residency, many systems | $50,000 and up | 4 to 6 months |
The cost ranges are typical planning figures. Partner companies on our AI chatbot development companies listing that name chatbots as a service reported averages from $250 to $25,000 in October 2026. Timelines are typical for one partner team. They stretch when your own team is slow to share data or approve answers.
AI Chatbot Development Cost by Type
Rule-based and template chatbots: $250 to $2,500
These bots follow a script. They show buttons, answer a fixed list of questions and hand the visitor to a person or a form when they get stuck. Many are built on a chatbot platform, not written from scratch. So the work is mostly setup, writing the replies and linking the bot to your website and email. Four of the 6 companies on our listing that name AI chatbot development as a service average between $250 and $2,500 per project.
This tier suits a small business that gets the same 20 questions every day. It won't answer anything outside the script. So it's a poor fit when customers ask open questions.
Custom AI chatbots that answer from your documents: $10,000 to $25,000
Most businesses mean this when they say "AI chatbot". The bot uses a large language model, but it answers from your own material: help articles, product sheets, policies, past tickets. The method is called retrieval-augmented generation, or RAG. The partner company builds a system that splits your documents into short passages and stores them in a search index. For each question, it finds the right passages. Then it asks the model to answer only from those.
The cost goes into that system and into testing. A good team writes a set of real customer questions with approved answers. It runs the bot against them before launch and fixes the misses. Our guide on how to evaluate a RAG chatbot proposal lists what that testing should include.
AI chatbots that take actions: $25,000 to $50,000
Here the bot doesn't just answer. It does things. It looks up an order, changes a delivery slot, raises a ticket, books a demo or updates a lead in the CRM. Each action needs a link to your system and rules for what the bot may and may not do. Anything that touches money or personal data needs a human check. Every connected system adds build and test time, so this tier costs more. Of the 6 companies on our listing that name AI chatbots as a service, one, Leadows Technologies, averages $25,000 per project.
If the bot plans several steps on its own, it's closer to an AI agent. Our guide to AI agent development cost covers that tier in more detail, and voice bots are covered in our AI voice agent cost guide.
Enterprise AI assistants: $50,000 and up
Large companies add needs that each carry their own cost. These include single sign-on and access rules, so staff only see what they're allowed to. They also include audit logs, several languages, data kept in one region and security reviews. The highest average project size among the chatbot partner companies on our listing is $25,000. Enterprise work often goes well beyond that once several teams and systems are involved.
What Drives the Cost of an AI Chatbot?
- Number of integrations. Each system the bot reads from or writes to, such as a CRM, ERP, helpdesk or payment gateway, adds design, build and test work.
- State of your content. Clean, current help articles are cheap to use. Scattered PDFs, old policies and answers that live only in people's heads take time to collect and fix.
- Channels. A website widget is the simplest. WhatsApp, Instagram, Slack, Teams and a mobile app each add setup and testing. Our WhatsApp chatbot development cost guide covers that channel's own fees.
- Languages. Every language needs its own test questions and review, even when the model handles the translation.
- Accuracy and safety bar. A bot giving product tips can tolerate small errors. A bot quoting prices, refund terms or medical information needs tighter testing, human review and logging.
- Analytics and handoff. A dashboard shows what people asked, where the bot failed and when it passed the chat to a person. It takes extra build time but pays for itself quickly.
Monthly Running Costs: The Part Quotes Often Leave Out
An AI chatbot pays the model provider for every message. Use is measured in tokens, and a token is about three-quarters of a word. Prices differ a lot between models. In October 2026, Anthropic lists Claude Haiku 5.5 at $0.10 per million input tokens and $0.50 per million output tokens. Claude Sonnet 5.5 costs $2 and $10, and Claude Opus 5.5 costs $4 and $20[1]. OpenAI publishes its model prices per million tokens on its own pricing page[2].
To show what that means, take a support chatbot handling 10,000 conversations a month. Each conversation has 6 exchanges. Each exchange sends about 3,000 tokens to the model: the question, the help passages it found and the instructions. It gets about 300 tokens back. That's 180 million input tokens and 18 million output tokens a month.
| Model | Input cost | Output cost | Monthly total |
|---|---|---|---|
| Claude Haiku 5.5 | 180M x $0.10 = $18 | 18M x $0.50 = $9 | about $27 |
| Claude Sonnet 5.5 with prompt caching | 120M cached x $0.10 + 60M x $2 = $132 | 18M x $10 = $180 | about $312 |
| Claude Sonnet 5.5 | 180M x $2 = $360 | 18M x $10 = $180 | about $540 |
| Claude Opus 5.5 | 180M x $4 = $720 | 18M x $20 = $360 | about $1,080 |
The caching line assumes 2,000 of the 3,000 input tokens are the same fixed instructions on every turn. Anthropic charges these at $0.10 per million on Sonnet 5.5 when they're read from cache[1]. Many teams also send simple questions to a small model and only hard ones to a larger model. That can cut the bill further.
On top of the model, budget for hosting the app and its search index, and for monitoring. Add a monthly support fee for updating content and fixing answers. Ask every partner company for a written running-cost estimate at your expected volume, with the model named, before you compare build prices.
AI Chatbot Development Cost in India vs the US
For US rates, the clearest public benchmark is the General Services Administration's contract price data. It lists the hourly ceiling rates firms have agreed with the US government[3]. We pulled the rates for the roles a chatbot project needs in October 2026.
| Role | Median GSA ceiling rate | Rates in sample |
|---|---|---|
| AI engineer | $175 an hour | 83 rates from 30 contractors |
| Machine Learning engineer | $170 an hour | 62 rates from 30 contractors |
| UX designer | $140 an hour | 297 rates from 143 contractors |
| QA engineer | $108 an hour | 125 rates from 57 contractors |
A mid-sized custom chatbot on your documents with one integration might need 320 hours of AI engineering, 60 hours of design and 80 hours of testing. At those medians, that's $55,955 plus $8,379 plus $8,634, or about $72,970. The same scope from an Indian partner company typically falls in the $10,000 to $25,000 range shown earlier. That is why many US and European companies hire Indian teams for this work. Our AI development cost guide compares rates across more AI roles.
A lower rate only saves money if the team delivers. Check that the people on your project have built and launched chatbots before. Make sure you'll talk to them directly, not through a sales contact.
A Worked Budget: Year One of a Support Chatbot
Here's how the numbers add up for one realistic project. It is a custom AI chatbot that answers from a company's help centre and order policies and hands hard cases to a person. It connects to one helpdesk and handles 10,000 conversations a month on Claude Sonnet 5.5 with prompt caching.
| Cost line | Built at US contract rates | Built by an Indian partner company |
|---|---|---|
| Build (first version) | $72,970 (460 hours at GSA medians) | $17,500 (midpoint of the $10,000 to $25,000 range above) |
| Model fees, 12 months | $3,744 ($312 a month) | $3,744 ($312 a month) |
| Support, 10 hours a month | $21,000 (120 hours at the $175 AI engineer median) | Ask for a monthly retainer quote |
| Year one before hosting | about $97,700 | about $21,200 plus support |
Two things stand out. Model fees are small next to the build and support when you pick a mid-sized model and cache the fixed instructions. But they'd nearly double on Claude Sonnet 5.5 without caching and more than triple on Claude Opus 5.5. Support is also a real line item. Answers go stale as prices, products and policies change. Someone has to keep the content and test questions up to date. Ask every partner company to price that retainer up front, alongside the build.
What AI Chatbot Partner Companies on UpBusiness Charge
We looked at the 12 partner companies on our AI chatbot development listing in October 2026. Only 6 of them name AI chatbot development as a service on their profiles, so the figures below cover those 6. Each passed our manual checks of registration documents, founder identity and three client projects. Their TrustScores range from 70.33 to 74.24 out of 100.
| Partner company | City | Average project size | Team | TrustScore |
|---|---|---|---|---|
| Leadows Technologies | Bengaluru | $25,000 | 11-50 | 71.33 |
| iTechCloud Solution | Surat | $10,000 | 11-50 | 70.33 |
| The Dreamer Designs India | Delhi | $2,500 | 11-50 | 73.67 |
| EXCELLIS IT | Kolkata | $1,000 | 51-200 | 74.24 |
| Propello AI Labs | Navi Mumbai | $1,000 | 1-10 | 70.33 |
| 10G SPECTRUM | Parakode | $250 | 1-10 | 70.33 |
The median average project size across the 6 is $1,750. Four average $2,500 or less: The Dreamer Designs India, EXCELLIS IT, Propello AI Labs and 10G SPECTRUM. They fit rule-based and template bots. iTechCloud Solution averages $10,000 and Leadows Technologies averages $25,000. Those sizes are closer to custom AI chatbots linked to business systems. Three of the 6 have 11 to 50 employees, so you'll usually work with a small senior team.
Average project size is the average of what each company reports across its client work, not a fixed price or a minimum. Use it to find companies whose usual work matches your budget, then ask each for a quote on the same brief.
Pricing Models for Chatbot Projects
| Model | How it works | Best for | Watch out for |
|---|---|---|---|
| Fixed price | One price for an agreed scope | Rule-based bots and clearly defined RAG bots | Change requests priced high once the scope is fixed |
| Time and materials | You pay for hours worked | Bots with several integrations or unclear requirements | Needs weekly reports of hours against progress |
| Fixed build plus monthly retainer | Fixed launch price, then a monthly fee for updates and support | Most custom AI chatbots | Check what the retainer covers and its response times |
| Platform subscription | Monthly fee to a chatbot platform, often per conversation or resolution | Teams that want a ready product, not custom software | Costs rise with volume and you can't take the bot with you |
For a custom AI chatbot, a fixed price for the first version followed by a monthly retainer works well for most buyers. It caps your launch cost. It also keeps someone in charge of answers that go stale. Our software development contract checklist lists the clauses to include on ownership, data and support.
Hidden Costs to Budget For
- Content clean-up. Someone on your side has to collect, update and approve the material the bot answers from.
- Test questions. Writing 100 to 300 real questions with approved answers takes your team's time, and the bot is only as good as that test set.
- Security. AI chatbots can be tricked into ignoring their instructions or leaking data through crafted prompts. The OWASP Top 10 for large language model applications lists prompt injection as the first risk[4]. Ask how the partner company tests for it.
- Model price changes. Providers change models and prices. Ask for a design that lets you switch models without a rebuild.
- Human handoff. If the bot passes chats to your team, you need staff, a helpdesk seat or a shared inbox to take them.
How to Reduce AI Chatbot Development Cost
- Start with one channel and the 20 to 50 questions that make up most of your volume. Add channels after launch.
- Fix your help content before the build starts, so the partner company isn't paid to work around gaps.
- Use a small model for simple questions and keep the larger model for the hard ones.
- Reuse a proven open framework for retrieval and chat rather than paying for custom plumbing.
- Agree a short paid pilot with a clear accuracy target before committing to the full build.
Rule-based or AI? If your customers ask the same questions in the same words, a rule-based bot is cheaper and more predictable. If they ask in their own words, about many products or policies, an AI chatbot is worth the extra cost. It handles questions a script can't. Our article on AI chatbots in customer service covers where each works well.
How to Get an Accurate Chatbot Quote
Vague briefs get vague prices. Before you ask for quotes, write down:
- The top 30 questions customers ask, and where the answers live today.
- Every system the bot must read from or update, with a contact who can grant access.
- Channels and languages for launch, and which ones can wait.
- Expected conversations per month, now and in a year.
- What the bot must never do, such as quote prices it can't confirm or change an order without approval.
Then ask each partner company for the same things. Ask for the model they'd use and why, and a monthly running-cost estimate at your volume. Ask for the testing plan, who owns the code and accounts, and support terms after launch. Lay the replies side by side using our guide to comparing software development proposals.
If you'd like quotes from verified partner companies, post your chatbot project on UpBusiness with your questions, systems, channels and budget, and we'll match you with three that fit. You can also browse verified AI chatbot development companies ranked by TrustScore.
