A chatbot is worth building when it captures something you are currently losing. If it isn't doing that, it's decoration with a monthly fee.
There is exactly one question that decides this: are you losing enquiries because nobody is available to answer them?
After hours, during jobs, on weekends, or simply because the volume is more than one person can handle. If yes, a bot has something real to catch and the payback is straightforward. If your phone gets answered promptly and your inbox is under control, a chatbot will mostly annoy people who would have converted anyway.
The useful ones are narrow. In practice a well-built bot:
Quote a firm price on a job it can't see. Promise a date the calendar can't honour. Make up a policy. Argue with an unhappy customer. These aren't theoretical — they're the failure modes that turn a bot into a liability, and defining the boundary is a real part of the build.
Roughly three tiers, and most small businesses belong in the middle one. A basic FAQ widget is nearly free and behaves like one. A proper lead-capture bot — trained on your business, wired into your calendar and CRM — is a low-four-figure one-time build. Operationally connected bots that read live inventory or pricing run higher.
The full cost breakdown with real numbers goes through the tiers, and what actually drives the price (hint: it's the integrations and the edge cases, not the conversation).
Write down your twenty most common customer questions. If you can do that in an afternoon, the build is fast and cheap because the hard part is already done. If you can't, start there — with a document, not a bot. That document is worth having whether or not you ever automate it.
A chatbot is also rarely the first automation worth buying. Faster lead response and follow-up usually return more for less. See business automation for how those compare.
The free audit looks at where your enquiries come from and whether a bot would actually catch anything.
Book a free audit