AI Website Chatbot Visitors get straight answers. Your team gets booked leads.

A chat assistant on your website, trained only on your services, policies and price list. It answers with a link to its source, books visitors who are ready and hands the rest to a person instead of guessing.

Free · 20 minutes · You talk to Syed · Scoped before anything is built

How it runs

AI Website Chatbot Running

A person steps in when needed

  1. A visitor asks a question

    On any page, at any hour, on a phone or a laptop.

  2. The assistant answers from your sources

    Only pages, policies and prices you approved, with a link to the one it used.

  3. Your tools update

    A booking, CRM lead or support ticket is created where your team already works.

  4. A person steps in when needed

    Hand-offs reach Slack, email or your help desk with the full conversation.

See it work

Watch it handle one.

A scripted walk-through of a typical website chatbot conversation — press play.

Simulated demo

AI website chatbot

Maple & Main Dental · Website chat New Patients page · Tue 9:14 pm · after hours
  1. Visitor

    Hi, do you direct-bill insurance? Mine is through work.

  2. AI assistant

    Hi, I’m Maple & Main’s virtual assistant; a person can take over any time. Yes, we direct-bill most workplace plans, so you pay only what your plan doesn’t cover.

  3. Visitor

    Good. Roughly how much are clear aligners with insurance?

  4. AI assistant

    That depends on your treatment and your plan, so I won’t guess. The dentist reviews costs at a consult, and monthly payment plans are available. Shall I book one?

  5. Answered from
    • Insurance & direct billing
    • Clear aligners
    • Payment plans
  6. Visitor

    Yes please. A weekday evening if you have one. I’m a new patient.

  7. Checking new-patient consult times… Thu 5:30 pm · Thu 6:15 pm · Tue 6:00 pm open Google Calendar
  8. AI assistant

    I have Thursday at 5:30 or 6:15 pm, or next Tuesday at 6:00 pm. Which suits you? I’ll also need your name and email for the confirmation.

  9. Visitor

    Thursday at 5:30. Dana Okafor, dana.okafor@example.com

  10. Consult booked Confirmed
    When
    Thu · 5:30 pm
    Visit
    New patient · aligner consult
    Cost question
    Handed to front desk
    Sources
    3 approved pages
  11. Logging the chat and handing off the cost question… Lead and summary in HubSpot · front desk notified in Slack HubSpot

Answered from 3 approved pages and booked in 1 min 52 s. The cost question went to a person.

Set this up for my business
What it replaces

Off your plate.

  • Answering the same ten questions by email every morning
  • Contact forms that wait unread until the next business day
  • Visitors leaving because they couldn’t find the policy
  • Copying chat and form details into the CRM by hand
How it works

Step by step.

  1. 01

    We gather what you’d want it to say

    We start with the 20 questions your team answers most and collect your pages, policies, price ranges and hours. You approve the source list; anything not on it is out of bounds.

  2. 02

    We build and test it on your real questions

    We match your tone and test it on real questions from your inbox, including the awkward ones. Each answer links to its source, or it says it doesn’t know and offers a person.

  3. 03

    It answers, captures and books

    When a visitor is ready, it asks your qualifying questions, books into your calendar or creates a CRM lead, and sends a confirmation. Your team gets a short summary.

  4. 04

    We review the gaps with you each month

    Every chat is logged. Each month we review unanswered questions and hand-offs with you, update the sources and retest as your prices and policies change.

Works with the tools you already use

  • Your existing website
  • Shopify
  • HubSpot
  • GoHighLevel
  • Google Calendar
  • Calendly
  • Zendesk
  • Freshdesk
  • Gorgias
  • Slack
  • Gmail / Outlook

Connected through official APIs where available. Names and logos are trademarks of their owners; no affiliation implied.

Guardrails

You stay in control.

It says it’s an AI

It introduces itself as a virtual assistant and offers a person at any point. Some US states, including California, require bots to disclose themselves; we do it everywhere.

It answers only from approved sources

In 2024 a Canadian tribunal held a business responsible for what its website chatbot told a customer. So it answers only from material you approve, links the source and gets a person when the material runs out.

No clinical, legal or financial advice

For clinics it sticks to scheduling, fees and policies; for firms, intake only. We minimize what chat collects, set a retention period and design around PIPEDA, PHIPA (vendors act as agents of the custodian) or HIPAA.

Money and exceptions go to a person

It can share a secure payment link from your own system, but never takes card details in chat. Refunds, discounts and price exceptions go to your team to decide.

In practice

What it looks like.

Example scenario · illustrative, not a client result

A three-dentist family clinic

Two people on the front desk. About 30 website questions a week about insurance, financing, hours and new-patient visits, plus 10 booking requests through the contact form. The assistant answers on the site and books new-patient consults.

About 2–2.5 hours a week back for the front desk (estimate)

The math
  • Questions the assistant answers fully: ~24/week × ~5 min each (find the policy, write the reply) ≈ 120 min/week
  • New-patient booking requests: ~10/week × ~4 min of back-and-forth ≈ 40 min/week
  • Hand-offs that arrive with details already collected: ~6/week × ~2 min saved ≈ 12 min/week
  • Total ≈ 172 min/week ≈ 2.9 h/week
  • Less 20% for overlap and ramp-up ≈ 2.3 h/week

Estimates use typical volumes and include a 20% allowance for overlap and ramp-up. Your audit measures the real numbers.

Why it matters

Your visitors ask what your team answers every day

Do you direct-bill insurance? Is there a call-out fee? What’s the cancellation policy? Most of these answers already live on your site, a few clicks deep. Visitors don’t dig. They send a contact form and wait, or try the next business that answers faster.

Meanwhile someone on your team spends part of every morning answering the same questions by email. The older button-menu chatbot rarely helps, because it can’t answer anything off script.

Answers with a source, or a person

The assistant reads only the material you approve: service pages, policies, FAQs and the price ranges you’re happy to share. Each answer shows the page it came from. If your material doesn’t cover a question, it says so and offers a person rather than guessing. A business is responsible for what its chatbot tells customers, so narrow and sourced is the point.

When a visitor is ready, it asks your qualifying questions, books into your calendar or creates a lead in your CRM, and sends a confirmation. Complaints, refunds and clinical or legal questions go to your team with the full conversation. Leads can flow into AI lead qualification, and the same approved material can power a private knowledge assistant for your staff.

Two useful variants

Support-ticket triage. In Zendesk, Freshdesk or Gorgias, it labels and prioritizes new tickets and drafts replies from your help articles for an agent to review.

Shopify order status. It answers “Where’s my order?” once the customer confirms their order number and email, and shares the tracking link. Returns and refunds still go to your team.

If your customers would rather message, the same playbook can run as a WhatsApp AI assistant.

How we roll it out

Audit. Build. Run.

Every engagement is scoped on a call. You’ll know the cost before anything is built.

  1. 01

    Audit

    • The 20 questions your team answers most, with today’s answers
    • Source list: the pages, policies and price ranges the assistant may use
    • Hand-off rules: what always goes to a person, and how it reaches them
  2. 02

    Build

    • Chat assistant on your site, styled to match and tested on real questions
    • Booking, CRM lead capture and hand-off alerts connected
    • Optional: help-desk ticket triage or Shopify order-status lookups
  3. 03

    Run & Optimize

    • Monthly review of unanswered questions and hand-offs
    • Source updates as prices, policies or services change
    • Plain-English report: chats, answers, bookings and hand-offs
FAQ

Questions owners ask about AI Website Chatbot

Last updated October 7, 2026

Can an AI website chatbot make up answers?

It’s built not to. It answers only from the pages and documents you approve, shows a link to the source, and when the material doesn’t cover a question it says so and offers a person. We test it on your real questions before launch and review the gaps every month.

Can it triage support tickets as well as answer sales questions?

Yes. It can label and prioritize new tickets in Zendesk, Freshdesk or Gorgias and draft replies from your help articles. An agent reviews and sends each draft until you decide which simple replies can go out on their own.

Can it tell Shopify customers where their order is?

Yes. It looks up the order once the customer confirms their order number and email, then shares the status and tracking link. Returns, refunds and address changes go to your team.

What happens to the leads it captures?

Each one lands in your CRM, such as HubSpot or GoHighLevel, with a chat summary and the answers to your qualifying questions, and your team is notified. Marketing emails or texts go only to people who agreed, in line with CASL in Canada and US consent rules.

What does an AI website chatbot cost to run?

It depends on chat volume, how much material it learns from and which tools it connects to. We scope it on a short call and give you a fixed build price and a clear monthly running cost before anything is built.

Let’s map AI Website Chatbot to your workflow.

Bring your call volume, your tools and the part that annoys you most. In 20 minutes you’ll know what it would take — and whether it’s worth doing.

Free · 20 minutes · You talk to Syed · Prefer email? muhammad@sadaat.ca

Talk about AI Website Chatbot