The chatbot price you pay is two separate costs: a one-time build and a monthly running cost that continues as long as the bot answers customers, and Benian Technologies, which scopes and builds business chatbots, tells buyers to price them separately. The build is driven by how many sources the bot must read, how many systems it must act in and how many ways a conversation can end with a person. The running cost is driven by conversation volume, how the vendor bills and the hours your team spends keeping the answers right.
That is why two quotes for the same chatbot can look nothing alike. One vendor shows a small monthly subscription and no build fee. Another shows a large build fee and a running cost that is mostly model usage. Neither number means much until you put both on the same timeline, usually the first year, and add the staff time each one assumes you will spend.
Benian publishes no price for any service, including the chatbot builds in its Chat AI work. This page explains how the industry bills, which choices move each line, and how to compare quotes, including when a subscription tool you set up yourself is the better buy.
How chatbot pricing models work: seats, conversations, resolutions and usage
Chatbot vendors bill in four common ways, and many combine two of them. Per seat pricing charges for each team member who logs into the inbox where the bot hands off conversations. It is predictable, but it charges you for your staff, not for what the bot does. Per conversation or per message pricing charges for traffic, so the bill rises in your busy season whether or not the bot helped.
Per resolution pricing charges only when the vendor counts a conversation as resolved by the bot. It sounds like paying for results, and sometimes it is. Read how the vendor defines a resolution. If a customer who gives up and leaves counts as resolved, you pay for a failure. Ask for the definition in writing and for a report that lists resolved conversations you can spot check.
Usage pricing is what a custom build mostly runs on. The language model provider charges by the amount of text processed, so long conversations and long documents pulled into each answer cost more than short ones. Hosting, a vector database for search over your documents and any messaging channel fees sit on top. That is why how much a chatbot costs per month is always a range tied to your traffic.
What moves the cost to build a chatbot
Chatbot development cost rises with four things. The first is sources. A bot that answers from one clean help center is a short project. A bot that answers from product pages, a returns policy, spec sheets in PDFs and a shipping table that changes weekly needs each source cleaned, kept current and tested against real questions.
The second is actions. Answering questions is the cheap part. Looking up an order, checking stock, booking a time or creating a ticket means connecting to your store, calendar or help desk, handling the cases where that system is slow or returns nothing, and deciding what the bot says when it cannot confirm.
The third is channels and languages. A website widget is one build. Adding SMS, WhatsApp or a second language multiplies the testing, because the same question has to come out right everywhere. The fourth is handoffs: where the bot stops and a person takes over, what context travels with the customer, and what happens outside staffed hours.
What moves the AI chatbot cost per month
Volume is the obvious driver, but answer length and retrieval matter almost as much. A bot that pulls five long document chunks into every reply processes far more text than one that pulls the single right paragraph. Tightening retrieval is often the cheapest way to lower a usage bill without making answers worse.
Model choice is the second lever. Larger models cost more per conversation and are not always needed. Many builds route simple questions such as hours or order status to a smaller model and keep the larger one for questions that need judgment. Test both on your own question log before deciding.
The third driver is upkeep, and it does not show up on any vendor invoice. Someone has to update the knowledge source when a policy, price or product changes, read a sample of conversations each week and fix the answers that went wrong. Budget that time as a named owner with a few hours a week, not as an afterthought.
Subscription chatbot or custom build, by situation
Pick a subscription tool, set up yourself, if your questions come from one help center, you need one channel, volume is modest and nobody needs the bot to take actions in other systems. Many help desk and e-commerce platforms include a chatbot that reads your existing articles. Turn it on, feed it your top questions and watch it for a month before paying anyone, including Benian.
A custom build starts to pay when the bot must read several sources, act in your own systems, follow rules a template tool cannot express, or run at a volume where per conversation or per resolution fees grow faster than usage costs would. It also matters when you want to own the work. Benian builds in accounts the client owns, with credentials the client holds, so the bot keeps running if the engagement ends.
Start smaller than you think in either case. Pick the twenty questions that make up most of your inbox, measure how often the bot answers them correctly, and expand only after that number holds. A bot that answers twenty questions well is worth more than one that answers two hundred badly.
Hidden chatbot costs and how to compare two quotes line by line
The costs that surprise buyers are wrong answers, review time and exit. A wrong answer about a return window or a product claim costs a refund, a complaint or worse, so the quote should say how the bot is tested before launch and what it does when it is unsure. Review time is the weekly hour someone spends reading transcripts. Exit cost is what you lose if you leave: the prompts, the cleaned knowledge base and the conversation history.
To build a fair chatbot pricing comparison, put each quote into the same five lines for the first twelve months: build or setup fee, platform or seat fees, usage or per conversation fees at your real volume, your team's upkeep hours, and what you keep if you cancel. Then ask each vendor three questions. What counts as a resolution or a billable conversation? What happens to my bill if volume doubles? Who owns the knowledge base and the conversation logs?
An online chatbot cost calculator can give you a rough starting range, but it cannot see your sources, systems or volume. Use it to sanity check a quote, not to replace one. If a quote you already have feels wrong, Benian's free Opportunity Map or a 30-minute call can tell you whether the scope matches the price before you sign.
What a storefront assistant handles in production
VOT Distribution, a multi-brand e-commerce distributor, runs two AI storefront assistants in production that Benian built as part of a wider growth system. The assistant at shopfreezo.com answers product, compliance and shipping questions around the clock.
Those three question types show where the build cost goes. Product answers need the catalog kept current. Compliance answers need careful wording, because a confident wrong answer is worse than a handoff. Shipping answers depend on rules that change. The linked case study shows VOT's client-reported results across the whole engagement, which includes outbound campaigns and content automation, so do not read those figures as the chatbot's results alone.