Skip to content

Booking by text, web chat or voice: how the options differ

Checked November 8, 2026

Which ways can an AI take a booking?

Three are common: text messages, a chat box on your website, and phone calls answered by a voice. A chatbot is software designed to converse through text or speech, so it can sit behind any of them1.

They don't work the same way, and they don't fail the same way. Here's how they compare.

Text message Web chat Phone voice
Where it happens The customer's phone Your website A phone call
What it needs A messaging number and consent rules A website that can host a chat box A phone line and speech recognition
Typical weak spot Short, unclear replies Customers who leave mid-chat Accents, noise, names and numbers
Hand-over to a person Staff reply by text Staff join the chat Transfer the call or take a message

How does booking by text work?

Some companies run their chatbots through messaging apps or plain SMS1. The customer texts your number and the tool replies.

Text suits quick, simple bookings. It also leaves a written record, which helps when someone says they were promised a different time.

What it needs:

  • a number customers can text
  • permission to message people. In the United States, the TCPA and the FCC's rules under it have a general rule that applies unless the recipient has given prior express consent2. Other countries have their own rules, so check your local regulator or the official source for your country
  • a way for staff to see the thread and reply

The risk is brevity. "Thursday?" could mean this Thursday or next. A good tool asks a follow-up and repeats the full date back before it books.

How does booking by web chat work?

A chat box sits on your site, and the customer types into it. It's the closest relative of the online booking form, but the customer can write freely.

You need a website you can add a chat box to, and the chat must be connected to the calendar. Without that link it can only answer questions, not book.

Chat is good when customers are already browsing your services and prices. It's weaker when they arrive wanting an immediate answer and nobody checks the chat until evening. Decide who watches it and when.

How does booking by phone voice work?

This is the hardest one. A voice system has to turn speech into words, work out what was meant, and answer aloud.

Older phone menus are called interactive voice response, or IVR. They answer callers, collect details and send calls on to the right person, using speech recognition and keypad presses3. An AI voice goes further by letting the caller speak freely.

Speech is messy. People differ in accent, dialect, pitch, volume and speed, and background noise and echoes distort the sound4. A salon with hair dryers running behind a caller, or a trade customer on a building site, is exactly that setting.

Names, street addresses and phone numbers are the weak points. Have the tool read each one back, and offer a text message to confirm the details in writing.

Voice has legal limits too, and they differ by country. In the United States, the FCC issued a ruling in February 2024 that speech made with generative audio counts as an "artificial" voice under the TCPA2. If your tool will ever place calls, read the FCC's own guidance or ask a lawyer first.

Choosing a channelSee how customers contact you nowPick one channel to startCheck what it needsTest it with staffAdd a second channel later
Start with the channel your customers already use.

Which one should a small business start with?

The one your customers already use. If they text now, start with text. If they browse and book on your site, start with chat. If your phone rings all day and you miss calls, voice may be worth the extra care.

Start with one channel. Test it. Add another later. Every channel needs the same calendar and the same booking rules, and each one needs a clear way to hand over to a person.

Cited material

  1. Chatbot, Wikipedia
  2. Telephone Consumer Protection Act of 1991, Wikipedia
  3. Interactive voice response, Wikipedia
  4. Speech recognition, Wikipedia