Short answer: if you get repetitive questions — hours, pricing, availability, "do you serve my area" — outside business hours or in volume, an AI chatbot usually pays for itself within a few months. If your site gets light traffic and one clear offer, it's often money spent on a feature nobody needed. Here's how to tell which one you are, what it actually costs in 2026, and what to avoid.
This isn't the chatbot you remember
The chat widgets that gave "AI chatbot" a bad name were rule-based: they matched keywords against a decision tree and broke the moment a customer phrased a question differently than the script expected. A 2026 chatbot is built on a large language model, reads your actual service pages, pricing, FAQs, and policies, and answers in natural language — while staying constrained to only talk about your business. The difference in practice is the gap between a customer typing "hours?" into a form that has no "hours" button and getting a dead end, and the same customer asking "are you open Sunday for a walk-in" and getting a real answer.
Where it earns its keep
After-hours FAQ deflection
If a meaningful share of your inquiries come in nights and weekends — common for contractors, salons, and service businesses — a chatbot answers the easy 80% (hours, pricing range, service area, booking process) instantly instead of the customer waiting until you're back online, at which point many have already messaged a competitor.
Lead qualification before a human gets involved
A well-built assistant can ask the two or three qualifying questions you'd ask anyway — project type, timeline, budget range, location — and hand you a structured lead instead of a one-line "how much do you charge?" email. This is the single highest-ROI use case for most of our clients.
Booking and scheduling assistance
Connected to a real calendar or booking system, the assistant can check availability and get someone onto your calendar without a back-and-forth email thread — genuinely useful for salons, consultants, and anyone who takes appointments.
Order status and account questions
For businesses with any kind of e-commerce or account system, "where's my order" and "how do I reset my password" are the highest-volume, lowest-value-per-answer questions your team handles. These are exactly what a chatbot should absorb first.
Where it's a waste of money
If your site gets a handful of visits a week, has one clear offer, and a phone number that's answered during business hours anyway, a chatbot has almost nothing to deflect and adds a recurring cost for a problem you don't have. The same is true for businesses whose questions are genuinely non-repetitive — every inquiry is different enough that a scripted or even LLM-assisted answer can't shortcut the conversation a human needs to have anyway. In both cases, the monthly platform fee is better spent on the things covered in SEO or local search that actually bring more of the right visitors to the site in the first place.
What it actually costs in 2026
Pricing splits into three rough tiers, depending on how much the assistant needs to do beyond answering questions from your existing content:
- Basic FAQ assistant (reads your site + a knowledge doc, answers questions, hands off to email/phone): $50–$200/mo platform cost, $500–$2,000 one-time setup
- Lead-qualifying assistant (structured questions, CRM/email integration, lead scoring): $150–$400/mo, $1,500–$4,000 setup
- Booking or account-connected assistant (live calendar, order lookup, account actions): usually bundled into a broader retainer rather than a flat fee, since it depends on what systems it has to talk to
The model/API usage cost itself (what the underlying LLM provider charges per conversation) is usually the smallest line item — for a typical small business volume it's single digits to low tens of dollars a month. Most of what you're paying for is the integration work: connecting it to your actual content, your calendar or CRM, and building in the guardrails below.
What to avoid
No human handoff
Any assistant that can't cleanly say "let me connect you with someone" and surface a phone number, email, or contact form when it doesn't know the answer will frustrate customers and cost you the lead. This is the single most common mistake we see in off-the-shelf chatbot setups.
Answering from nothing
An assistant that isn't actually grounded in your real content will confidently make things up — wrong hours, wrong pricing, promises you didn't make. It needs to be restricted to answer only from your actual site content and documents, not from general knowledge about your industry.
Aggressive, unskippable popups
A chat window that auto-opens and blocks the page the moment someone lands is a conversion killer, independent of how good the AI behind it is. Let it sit as a small, dismissible affordance in the corner.
No monitoring
Someone needs to periodically read real conversation logs. Assistants drift, edge cases show up, and the only way to catch a wrong answer before it costs you a customer is to actually look.
A quick readiness check
You're a good candidate if most of these are true: you get inbound questions outside business hours, the same three or four questions come up repeatedly, you have real site content and FAQs for the assistant to ground itself in, and you have a real fallback (phone, email, a human) for it to hand off to. If two or fewer of those are true, it's worth fixing the site's content and search visibility first — an assistant with nothing to answer from and nothing answering it doesn't help.
We build AI assistants grounded in your actual site content and connected to your real booking or CRM tools — not a generic bolt-on widget.