"How much does a chatbot cost?" is a bit like asking how much a website costs. A simple bot that answers ten common questions is a very different project from an assistant that checks orders, books appointments and hands complex cases to your team. The good news is that the price follows a handful of clear drivers, and once you know them you can control the budget.
One-off costs and ongoing costs
Split the budget into two parts before comparing quotes. One-off costs cover planning, building, connecting and testing the bot. Ongoing costs cover the AI model's usage, hosting, monitoring and improvements after launch. A quote that only mentions the build is only half the picture.
What drives the build cost
- Scope: the number of jobs the bot does. Answering FAQs is one job; qualifying leads, booking calls and checking order status are three more.
- Integrations: every connection to a CRM, calendar, shop, helpdesk or database adds design and testing work.
- Knowledge: if answers must come from your own documents, those documents need collecting, cleaning and organising first.
- Channels: a website widget is the simplest; WhatsApp, Instagram, email and voice each add work.
- Languages: supporting more than one language means more testing and content checks.
- Handoff to people: deciding when and how the bot passes a conversation to your team, with the history attached.
- Testing and safeguards: building a set of real questions, checking answers, and preventing the bot from going beyond what it should do.
What drives the running cost
- Model usage: AI providers usually bill by usage, so cost rises with the number and length of conversations. Providers publish their current prices.
- Hosting and storage for the bot and any searchable knowledge base.
- Monitoring: reviewing conversations, catching wrong answers and tracking how often people need a human.
- Updates: new products, policies or prices mean the knowledge base needs refreshing.
- Platform fees if the bot runs inside a third-party tool.
Ways to keep the cost down
- Start with the questions you get most often, and leave rare cases to your team.
- Connect one system first, usually your calendar or CRM, and add others once the first works.
- Clean up your documents before the build; poor content is the most expensive thing to fix later.
- Choose a model that is good enough for the task rather than the largest one available.
- Agree how success will be measured, such as questions resolved and handoffs, so you only pay to improve what matters.
Questions to ask before you get a quote
- Which tasks must the bot handle on day one, and which can wait?
- Which systems must it read from or write to?
- Where do the answers come from, and who keeps that content up to date?
- What should happen when the bot doesn't know the answer?
- Who will review conversations after launch, and how often?
What AgenCoder charges
Our projects start from $3,000. The final price depends on the scope of work: the tasks, integrations, channels and testing involved, and whether you want ongoing support. After a short discovery call we send a written proposal with what's included, so you can compare options properly.
Be wary of any quote that doesn't mention testing or what happens when the bot is wrong. Those are where cheap chatbots fail, and where customers lose trust.
