AI receptionist basics
What is an AI receptionist? A plain guide for small businesses
What an AI receptionist does, where it helps a small business, what it should never do on its own, and how to tell a good one from a chatbot.
The short answer
An AI receptionist is software that handles the first conversation with a customer the way a good front desk would. It answers the common questions, asks what you would ask, takes contact details, books or requests an appointment, and passes anything else to a person with the conversation written up.
It usually works on your website first, as chat or voice, and then on messaging channels such as WhatsApp and email. Some products also answer phone calls.
The difference from a chatbot is where the conversation ends. A chatbot answers a question and stops. A receptionist moves the enquiry to its next step and makes sure somebody owns it.
What a good AI receptionist does
The work is the same as at a front desk. What changes is that the software does it at any hour, and consistently.
- Answers questions about your hours, services, prices and policies from information you approved, and can show where each answer came from.
- Asks the questions you would ask before taking on the work, such as which service, whether it is a first visit, or the budget and timeline.
- Takes the customer's name and a way to reach them, and writes the enquiry up so nobody has to call back for missing details.
- Books real free times when it is connected to your scheduler, or takes a request for a person to confirm when it is not.
- Knows your opening hours and says plainly when you are closed and when you next open.
- Hands the conversation to a named person when it needs judgement, with the transcript attached.
If a product cannot do the last two, it is a chatbot with a receptionist's name.
What it should never do on its own
The risk with any automated front desk is confidence without authority. A well-built receptionist has a written boundary, and it keeps to it.
- Give medical, legal, financial or other professional advice.
- Invent a price, a discount or a policy that is not in your information.
- Say an appointment is booked, a refund approved or a record changed when it has only taken a request.
- Promise a callback time nobody on your team can meet.
- Keep a customer going round in circles after they ask for a person.
Helm publishes its boundary on what Helm handles, and you add your own rules for what must go to a person. Whichever product you choose, ask to see its boundary written down.
AI receptionist, chatbot or answering service?
The three are easy to confuse because they can all say hello on your behalf. They differ in what happens next.
| Website chatbot | Answering service | AI receptionist | |
|---|---|---|---|
| Answers common questions | Yes, from a FAQ or your website | Yes, from a script | Yes, from approved information, with sources |
| Takes a complete enquiry | Sometimes, as a form | A message, sometimes intake questions | Yes, with the answers to your questions |
| Books appointments | Rarely | Some services, into your calendar | Where a scheduler is connected, otherwise a request |
| Out of hours | Always on | Depends on the plan | Usually always on |
| Usual pricing | Monthly | Per minute or per call | Monthly, often with an allowance of minutes or conversations |
The fuller comparison, including when a human answering service is the better choice, is in AI receptionist vs answering service.
Where it helps most
An AI receptionist earns its place when enquiries repeat and arrive while the team is busy. That describes most service businesses that take appointments or quote for work.
- Clinics and practices, where reception spends the day on prices, insurers, parking and cancellations.
- Salons, spas and studios, where messages arrive while everyone is with a client.
- Consultants, agencies and advisers, where first calls are spent on questions the website could answer.
Some signs you would notice: messages answered hours late, the same five questions every day, enquiries arriving in the evening and at weekends, and call-backs made only to collect details that should have been asked the first time.
How to judge one before you pay
Demos are built to go well. Test on your own information, and spend most of the time on the questions that should not go well.
- Build it from your own website and ask the ten questions your customers ask most.
- Ask something that is not in your information. A good receptionist says it does not know and offers to pass it on.
- Message it outside your hours. It should say you are closed before it promises anything.
- Ask for a person. It should take your details and hand over, not argue.
- Look at the handoff your team receives. It should read like a complete enquiry, not a transcript to decode.
- Try to book with no scheduler connected. It must not say the appointment is confirmed.
Then check the pricing model against your volume. The cost guide explains the models you will meet.
How Helm approaches it
Helm builds a preview receptionist from your website in about a minute, and you can test it on chat for free. Website chat is available today, website voice comes with an add-on, and WhatsApp, email and a phone line are in early access, set up with you.
Going live costs $500 a month on Core for one receptionist, or $700 a month on Scale for up to 5. Full details are on the pricing section.
FAQ
Frequently asked questions
Is an AI receptionist the same as a virtual receptionist?
Not quite. A virtual receptionist usually means a person working remotely, often through an answering service. An AI receptionist is software. Some services combine the two, with software handling the first conversation and people taking over.
Can an AI receptionist answer phone calls?
Some products answer calls directly. Helm answers on website chat today, and by voice on your website with the Voice: website + phone add-on; a phone line is in early access as part of the same add-on, set up with you.
Will customers know they are talking to AI?
They should. Say it is an assistant in the greeting, answer accurately, and make reaching a person easy. Customers are far less forgiving of being misled than of talking to software.
How long does it take to set up?
A first preview from your website takes about a minute. Expect a first full setup to take 20 to 30 minutes, most of it deciding your services, hours and intake questions rather than typing.