How to Add an AI Chatbot to Acuity Scheduling (3 Ways)

No. Acuity Scheduling (formerly Squarespace Scheduling) gives clients a scheduling page, confirmation emails, and reminders — there is no conversational layer. Every question the scheduling page doesn't answer — “do you do color corrections?”, “what should I bring?”, “is there parking?” — becomes an email, an Instagram DM, or a phone call that a human answers by hand, one at a time.

The gap extends past pre-booking questions. Acuity's client self-service is deliberately narrow: clients can reschedule only through the link in their confirmation email, one appointment at a time, date and time only. When that email is buried or deleted, the request lands with your front desk too. A chatbot done right absorbs both streams — questions before the booking and changes after it.

Before comparing approaches, be clear about the job. “Chatbot” covers everything from a glorified FAQ page to an agent that completes a paid booking. For an appointment business on Acuity, the bar looks like this:

• Service and pricing questions, answered from your actual catalog — not from a model's general knowledge of what businesses like yours usually charge.
• Availability questions (“anything Thursday after 4?”) answered from the live calendar, not a stale copy.
• Policy questions — cancellation windows, deposits, prep instructions — with answers consistent with what you actually enforce.
• The booking itself: pick a real open slot, collect client details, take payment, confirm.
• Post-booking changes: route reschedule and cancellation requests somewhere self-serve instead of ending the conversation with “please call us.”
• Escalation: recognize when a human is needed and get out of the way.

Judge every option below against this list. Most setups fail on grounding, on booking, or on both.

The fastest route to something on your website is assembling a chain: a chatbot platform for the widget and the conversation, plus a Zapier-style automation tool as glue between the bot and Acuity. The build looks like this:

1. Pick a chatbot builder that offers an embeddable website widget.
2. Type your services, prices, and policies into the bot's knowledge base or conversation flows — manually.
3. Connect an automation platform to Acuity so the bot can trigger actions like creating an appointment or looking up a client.
4. Map the booking intent either to a link out to your Acuity scheduling page, or to a multi-step automation that creates the appointment.
5. Test every path — then re-test the whole chain every time you change a service, a price, or a policy.

The strengths are real: no developer, live in days, and enormous flexibility in what you can wire together. The weaknesses are structural. The bot's answers are grounded in whatever you typed in, so the moment you raise a price or add a service in Acuity, the bot is confidently wrong until someone remembers to update it. True in-chat booking with payment is rare — most chains punt to a scheduling-page link, at which point your “booking chatbot” is a FAQ widget with a link on the end. And every hop between tools is a failure point that tends to fail silently.

If you have engineering resources, you can build exactly what you want: a backend that calls the OpenAI API with your own system prompt, tools that read services and availability from the Acuity API, and a chat widget on your site. The shape of the project:

1. Stand up a backend service that holds your OpenAI key and system prompt and streams responses to a chat widget.
2. Give the model tools (function calls) that query the Acuity API for services, prices, and open slots.
3. Build the booking flow — slot selection, client details, payment collection — and the error handling around it.
4. Build the embeddable widget, secure it against abuse, and add logging so you can see what the bot actually says to clients.
5. Maintain all of it: model deprecations, Acuity API changes, prompt regressions, security patches.

This is the only approach with no ceiling — you own every behavior. It's also the only approach where you own every failure. Expect weeks of developer time before the edge cases are handled, and a permanent line item for maintenance. For most appointment businesses, the honest question isn't “can we build this?” but “is this what we want to be maintaining in two years?”

Tecton Scheduler takes a different route: it layers on top of your existing Acuity account — it doesn't replace it — and keeps a synced copy of your services and appointments, on demand or on a schedule, with credentials encrypted at rest. Its AI booking assistant is grounded in that synced catalog plus documents you upload, so answers reflect what's actually on your calendar. And when the conversation reaches “book it,” the assistant hands off to a real booking flow with live Acuity availability and Square payment — not a link out.

1. Connect your Acuity account. Tecton syncs your services and appointments automatically.
2. Add your own OpenAI API key, and choose the model and system prompt — the assistant runs on your key, under your rules.
3. Upload supporting documents: policies, FAQs, prep instructions, anything the catalog doesn't cover.
4. Copy one script tag onto your website — that's the entire embed.
5. Set origin restrictions so the widget only loads on the domains you allow.

Bookings made through the assistant come from live Acuity availability, so double-booking isn't a failure mode you have to engineer around. Afterward, the activity flows into the rest of the layer: revenue and booking analytics with prior-period comparisons, and a read-tracked notifications feed covering reschedules, fees charged, and declined cards. Plans (Starter, Growth, and Scale) are priced to booking volume, month-to-month, with a free trial on every tier.

The pattern: no-code buys speed at the price of grounding and reliability; custom buys control at the price of ownership; a native assistant buys grounding and real booking at the price of adopting a platform. If the chatbot's job is to answer accurately and put paid bookings on your Acuity calendar, the integration is the product — the chat window is the easy part.

A chatbot is the front door, not the whole house. The same clients who ask questions before booking will want to change appointments after — and Acuity's built-in answer (the confirmation-email link) is exactly the dead end the chatbot was supposed to remove. Tecton pairs the assistant with a self-serve reschedule page: clients verify their email and phone, pick a new slot from live Acuity availability, and pay any reschedule fee via Square or Stripe at confirmation. The full setup is covered in how to let clients reschedule in Acuity.

For agencies and multi-location operators, each business runs its own assistant grounded in its own synced catalog, managed from a super-admin console with per-tenant feature flags, usage, and billing monitoring. Since Acuity's official answer to running multiple businesses is separate accounts, that layer matters — see running Acuity across multiple locations.