AI receptionist for medical practice

It's 7.50 on a Monday morning and the lines open at 8. By 8.05 your two receptionists are each on a call, four more people are queuing, and the practice manager is answering the phone in her office because the HGV driver who needs a medical before Friday won't stop ringing. Most of those calls are not clinical. They're "can I book a blood test", "how much is a private GP appointment", "do you do travel jabs" and "where do I park".

An AI receptionist for medical practice takes that layer of admin calls and bookings so your people can deal with the patients who need a person. This page is about where it fits, where it doesn't, and what it must never do. That last part comes first, because it's the part that matters most in a medical setting.

Who an AI receptionist for medical practice suits

Private practices are the natural fit, and we'd rather say that up front than discover it in your procurement meeting. A private GP clinic, a menopause or men's health clinic, an occupational health provider doing pre-employment and driver medicals, a travel clinic, a private specialist's rooms: these run their own booking, set their own prices, choose their own phone supplier and answer to the CQC and the ICO, not to an NHS contract.

NHS GP surgeries are different. An NHS practice works under the GP contract, buys telephony through a national framework, reports call data to NHS England and has clinical safety duties for any digital tool it adopts. Those duties apply to every AI phone product, not only this one. We cover them in detail further down, in the section on the GP practice AI phone system rules. The short version: an NHS practice must check those requirements, with its ICB, before putting any AI tool on its phone line, and Klugon isn't an NHS-assured supplier today.

So read the rest of this page with a private clinic in mind unless a section says otherwise.

The calls a medical practice actually gets

Medical reception is not one kind of call. It's a dozen kinds arriving at once, and the skill of a good receptionist is sorting them fast. Here's how they tend to break down in a private clinic.

Booking calls by appointment type

Private GP consultations, usually 15, 20 or 30 minutes, often with a choice of doctor.
Nurse appointments: cervical screening, blood pressure checks, wound dressings, ear irrigation where the clinic offers it.
Blood tests, many of them fasting, which means early-morning slots only.
Travel health consultations and vaccinations, booked weeks ahead of a trip and often in courses that need a second or third dose.
Medicals: HGV and PCV driver medicals (D4 forms), taxi licence medicals, pre-employment, insurance, diving and sports medicals.
Health screens, from a basic well-person check to a long executive screen that needs a double slot and a fasting blood draw.
Specialist clinics you run on set days: menopause, sexual health, minor surgery, joint injections, dermatology.

Each of those has its own length, its own clinician and its own preparation rules. A receptionist holds all of that in their head. The AI holds it in your service list and your written instructions, which is why setup matters more than anything else on this page.

Admin and information calls

A large share of calls don't end in a booking at all. Callers want your prices, your opening hours, whether you take their insurer, what to bring to a driver medical, whether they need a GP referral to see a specialist and how long a report takes to come back. Patients ring to ask if you can see children, if there's step-free access, and whether the travel clinic stocks a particular vaccine.

These are the calls an AI answers well, because the answers are written down and don't change much from week to week.

Calls that are clinical, sensitive or both

Then there are the calls that need a clinician or an experienced member of staff: test results, medication questions, "my child has a rash, should I bring her in", repeat prescriptions, complaints, safeguarding concerns, a relative asking about another patient's care, a distressed caller. We deal with these in their own section. The rule is simple. The AI doesn't handle them; it hands them over.

When the phones ring hardest

The busy pattern in a medical practice is well known to anyone who has worked reception. Monday morning is the peak of the week, especially after a bank holiday. The first hour after the lines open is the busiest of any day. Lunchtime brings a second smaller wave from people calling on their break, and late afternoon brings the "I've just finished work" calls, which in a private clinic often go to voicemail because reception has gone home.

The year has a shape too. Autumn brings flu and other seasonal vaccine queries from September. Travel clinics get busy from January, when people book summer holidays, and again before Easter and the long summer break. Health screens pick up in January as part of New Year resolutions. Occupational health providers see bursts when an employer onboards a group or when a batch of driver licences comes up for renewal. For a private clinic, the evening and weekend gap is where calls are lost most often: a prospective patient who rings a private GP at 8pm on a Sunday will usually try the next clinic on the list if no one answers.

You can see your own pattern in Klugon's analytics once calls start coming in, by hour of day and day of week.

What it handles for a medical practice

Klugon's AI phone receptionist answers every call when your plan is live, around the clock, with no business-hours switch. Here's what it can do in a clinic and how to set the rules.

Booking with your clinic's rules

The AI books into Klugon's built-in booking calendar. That calendar uses your opening hours and timezone, the length of each service, the staff who deliver it and a few rules you control:

Slot interval: 10, 15, 20, 30, 45, 60, 90 or 120 minutes. A clinic with 15-minute GP slots and 30-minute nurse slots sets service lengths to match.
Buffer after bookings, useful after a long health screen or minor surgery so the room can be turned round.
Minimum notice, so no one books a slot that starts in 20 minutes while the doctor is already behind.
How far ahead patients can book, for example 60 days, which stops travel jab courses being booked six months out before stock is confirmed.
Closed dates, for bank holidays and training afternoons.
Staff assignment, so a smear test is only offered with a nurse who does them, and the HGV medical only with a doctor who signs D4 forms. Conflicts are checked per clinician.

The patient gets a booking reference in the same conversation, and the booking shows up in your dashboard, where staff can edit, move or cancel it. On the phone the AI can cancel a booking for a caller; rescheduling is done by cancelling and booking again, or by your staff. In chat and messaging it can reschedule directly.

Rules that the calendar can't express as a setting go in your written instructions. "Fasting blood tests are only booked before 10am." "Second doses of the rabies course must be seven days after the first." "Children under 2 are seen by Dr Ahmed only." The AI follows these the way a new receptionist follows a crib sheet. Test them before you go live, because it's following instructions, not reasoning clinically.

One honest limit: Klugon doesn't connect to clinical systems such as EMIS, SystmOne or the private practice systems clinics use. If your diary lives in one of those, Klugon's calendar is a second diary, and someone has to keep the two in step. Some clinics give the AI a protected set of slots (new patient consultations, or medicals) and leave everything else in their main system. After setup, the AI can also check busy times in Google, Outlook or Apple calendars and send a Calendly booking link, on the Business tier or above.

Questions it answers from your own information

You give the AI your fee list, your policies and your practical details by uploading documents (PDF, Word, text or CSV, up to 20 MB each), pasting text, or scanning your website. It answers from that knowledge and hands over when it doesn't know. Typical clinic questions:

What a private GP appointment costs, and what's included (for example, whether a follow-up within 14 days is free).
Which insurers you're recognised by, and what the patient needs from their insurer before booking, such as a pre-authorisation number.
What to bring to a driver medical: photo ID, glasses or contact lenses, the D4 form, a list of medications.
How to prepare for a fasting blood test.
Where to park, which entrance has step-free access, and whether there's a lift.
Your cancellation policy, as you've written it. The AI can tell callers what it is. It can't charge a fee or take a deposit.
How long reports and letters usually take.

It answers calls in English, German, French, Spanish, Italian, Portuguese, Dutch, Japanese and Arabic. That's useful for clinics with international patients, but don't let it stand in for a professional interpreter in anything clinical.

Calls it must hand over to a person

This is the most important section for any clinic. A medical practice that uses an AI receptionist carelessly can put a patient at risk. Here's how to set it up so it can't make a clinical judgement.

Emergencies and symptoms: no triage, ever

The AI must not decide whether a symptom is serious. Not "that sounds like it can wait", not "I'd suggest an urgent appointment", and not a list of red-flag questions. Triage is clinical work, and the CQC treats AI triage tools as a distinct category with their own regulatory requirements.

What you set instead is a fixed message it reads whenever a caller describes symptoms or asks for medical advice, for example:

"I'm the clinic's automated assistant and I can't give medical advice. If this is an emergency, please hang up and call 999. If you need medical help quickly and it isn't an emergency, call NHS 111 or go to 111 online. I can book you an appointment or pass a message to the team."

The same applies in chat and on WhatsApp. For patients in the US or Canada, the emergency number is 911, and you'd write the message accordingly.

Then it flags the conversation for your team. It doesn't hold the patient on the line asking questions to work out urgency. It doesn't book the "next available" slot as a substitute for medical help.

Calls that always go to staff

Write these into your instructions as "always hand over":

Test results of any kind, including "has my result come back yet".
Medication: doses, side effects, interactions, repeat prescriptions, "can I take this with that".
Anything about another patient, including a parent asking about an adult child, or an employer asking whether an employee attended a medical. Confidentiality decisions belong to staff.
Safeguarding concerns and anything that suggests a vulnerable person is at risk.
Complaints and clinical concerns about care already received.
Distressed callers, and anyone who mentions self-harm. The AI reads the 999 and 111 message and flags the conversation.
Fitness-to-drive questions, such as "will my condition stop me passing the HGV medical". That's a conversation for the doctor.
Requests for records or reports, which usually have identity checks and fees you manage.

What "handing over" means in Klugon

Be clear on the mechanics before you rely on them. When the AI hands a conversation to your team, it marks the conversation as needing a person, logs it in the inbox with a "Needs a person" banner and sends a notification to Slack or Zapier if you've connected one. The insights page lists handovers that are still waiting.

It doesn't put the caller through to a person. There's no live call transfer. The caller is told the team will get back to them, and a member of staff calls them back from your normal phone. On WhatsApp, Instagram and Messenger, staff can reply from the Klugon inbox within Meta's 24-hour window. Phone and website-chat conversations can't be replied to from the inbox, so staff phone or email the patient themselves.

Two more limits to know. The AI isn't paused after a handover; if the patient keeps talking, it keeps replying, so your instructions should tell it to stop booking and keep repeating that the team will be in touch. And Klugon doesn't send email or text alerts, so someone needs to watch Slack or the inbox during opening hours, with a plan for handovers that arrive overnight.

Sample conversations

These are samples written for this page. The clinic is made up, and the AI voice on Klugon's site is AI-generated. Nothing below is a recording.

Sample call: booking a menopause consultation

A sample call. The business, Harbourside Private GP in Bristol, is made up.

AI: Thank you for calling Harbourside Private GP. I'm the clinic's automated assistant. How can I help?

Caller: Hi, I wanted to book with the menopause doctor. A friend said you have a clinic for it.

AI: We do. Dr Okonjo runs the menopause clinic on Tuesday and Thursday afternoons. A first consultation is 40 minutes and costs £195. Would you like me to look for a slot?

Caller: Yes please. Thursdays are better.

AI: The next Thursday slots are 2.20pm on the 6th and 3.40pm on the 13th. Which would you prefer?

Caller: The 13th.

AI: Can I take your full name and a phone number?

Caller: Joanne Pike, 07700 900461.

AI: Thanks, Joanne. You're booked with Dr Okonjo on Thursday the 13th at 3.40pm. Your reference is HPG-4471. The clinic asks you to bring a list of any medicines or supplements you take. Is there anything else?

Caller: Should I stop my HRT patches before I come?

AI: I can't give medical advice, so please don't change anything based on what I say. I'll pass your question to the clinical team so someone can call you back. If you feel unwell and need help quickly, call NHS 111.

Caller: Okay, thanks.

The question about patches is flagged for the team in the inbox and in Slack. The booking itself is done.

Sample call: a caller with chest pain

A sample call. The business is made up. This shows the fixed safety message, not a clinical decision.

AI: Thank you for calling Harbourside Private GP. I'm the clinic's automated assistant. How can I help?

Caller: I've had a tight chest since this morning and it's going into my arm. Can I see a doctor today?

AI: I can't give medical advice. If this is an emergency, please hang up now and call 999. If you need medical help quickly and it isn't an emergency, call NHS 111. I'm letting the team know you called.

Caller: I don't want to make a fuss.

AI: Please call 999 now. I'm an automated assistant and I can't help with symptoms. The team will see your message.

The AI didn't ask questions to grade the pain, didn't offer an appointment and didn't reassure. That's the behaviour you want, and it's why the safety message should be fixed wording you test, not something the AI composes each time.

Sample website chat: an HGV medical

A sample chat on the clinic's website. The business is made up.

Visitor: do you do HGV medicals and how much

AI: Yes. HGV and PCV driver medicals are £75 with Dr Lewis or Dr Shah, and appointments are 30 minutes. Please bring your D4 form, photo ID, your glasses or contact lenses if you wear them, and a list of your medicines. Would you like to book?

Visitor: yes, need it before friday, my licence runs out

AI: The earliest slots are Wednesday at 8.30am or Thursday at 12.15pm. Which suits you?

Visitor: wednesday

AI: Done. You're booked for Wednesday at 8.30am with Dr Shah. Your reference is HPG-4502. Can I take your name and mobile number for the booking?

Visitor: Darren Kell 07700 900218. will having diabetes be a problem

AI: That's a question for the doctor, so I won't guess. I've passed it to the team, and they'll contact you before your appointment if they need anything. Bring details of your diabetes treatment with you.

GP practice AI phone system: what NHS surgeries must check first

If you run or manage an NHS GP practice in England, this section is for you, and it's deliberately blunt. A GP practice AI phone system isn't a decision a practice makes on price and features alone. Before any AI tool touches your patient line, including Klugon, check these points with your ICB and your clinical safety lead.

NHS England's advice on virtual (AI) receptionists

In February 2026 NHS England issued advice to practices about virtual (AI) receptionist products, circulated by NHS North East London ICB. It says the decision rests with the practice, but asks practices to do their own assurance before going ahead. The points it raises:

Telephony procurement. The GP contract requires practices to buy telephony from the Better Purchasing framework once their current contracts expire, and the advice notes that most virtual receptionist products are not supplied by a framework supplier. Klugon is not a Better Purchasing framework supplier.
National reporting. Practices provide call data on set metrics through a national extraction. If calls are routed through an AI product first, the practice and its telephony supplier may lose sight of total call volumes. The advice asks practices to get assurance on how that reporting will still be delivered.
Porting and diversion. NHS England's programme view is that practices should think carefully before porting the practice number to a third party, and that forwarding calls to a third-party service may bring extra charges from the telephony supplier. Klugon's "keep your number" option works by forwarding your calls to a Klugon number, which is exactly the arrangement this point is about.
Clinical safety and information governance. Practices must carry out the required clinical safety assurance against DTAC and DCB0129, and handle patient data in line with information governance and cyber security standards, which the advice says include ISO 27001 and Cyber Essentials. Klugon doesn't hold ISO 27001 certification, and we can't hand you a DTAC submission or a DCB0129 clinical safety case today.
Continuity. Practices should agree what happens if the AI system fails, and how support is split between two suppliers.
Other routes. Patients who struggle with automated systems must still be able to reach the practice another way, and an AI receptionist should add to access, not replace online consultation or ordinary phone access.

Clinical safety standards and the CQC

Two NHS clinical risk management standards apply. DCB0129 is for the developer; DCB0160 is for the organisation deploying the technology, and NHS England Digital describes compliance as mandatory under the Health and Social Care Act 2012. The CQC's GP mythbuster 109 on AI in GP services, published in July 2025, says adopters must nominate a clinical safety officer who is a senior registered clinician, keep a hazard log and risk assessments, and procure AI tools in line with DCB0160, DTAC and MHRA rules where they apply. The CQC says it will look for this evidence at assessment.

Data security

Any organisation with access to NHS patient data and systems completes the Data Security and Protection Toolkit. Adding a supplier that handles patient conversations affects your DSPT answers and your data protection impact assessment.

Our honest view for NHS practices

For the main patient line of an NHS GP surgery, the checks above are significant, and Klugon doesn't currently carry the NHS assurance documents that would make them straightforward. We'd rather you knew that now. A practice that runs a separate private service, such as a private travel clinic, should still take advice from its ICB before using any AI tool, because the line between the NHS and private sides is rarely clean on the phone.

Setup checklist for a private clinic

The work is mostly writing things down. Plan a few hours with your practice manager and a clinician, then a week of testing in the browser during the free trial before any patient hears it.

Decide the scope. Which lines and channels: the main number, after-hours only, the website chat, WhatsApp. Many clinics start with evenings and weekends.
Build the service list. Every bookable appointment type with its length, price and the clinicians who deliver it. Be specific: "Private GP 20 min", "Fasting blood test 10 min, before 10am", "HGV medical 30 min, Dr Lewis or Dr Shah".
Set opening hours, slot interval, buffers, minimum notice, how far ahead and closed dates. Add bank holidays and training afternoons now.
Write the safety message. Fixed wording, with 999 and NHS 111 (or 911 for North American patients). Have a clinician approve it.
Write the hand-over list. Results, medication, third-party enquiries, safeguarding, complaints, distress, fitness-to-drive. Tell the AI to stop booking once it hands over.
Load your knowledge. Fee list, insurer list, preparation instructions, cancellation policy, parking and access, report turnaround times. Upload them as documents or paste them in. Remember the AI answers from this, so out-of-date prices produce out-of-date answers.
Write the greeting. Klugon's default greeting doesn't say it's an AI, so write your own that does, for example "I'm the clinic's automated assistant".
Connect Slack or Zapier so handovers and bookings reach whoever covers reception. Agree who checks overnight handovers first thing.
Test the hard calls. Ring it in the browser and describe chest pain, ask about a result, pretend to be a worried parent, ask about someone else's appointment. Check that the safety message comes out word for word and that each one is flagged.
Do your data protection paperwork. Update your privacy notice and DPIA, and record Klugon as a processor (see the compliance section).
Go live after the trial. A phone number can only be added once the 7-day trial ends. You either buy a new number or forward your existing number to a Klugon number using your phone provider's call forwarding. Our guide to keeping your number with call forwarding walks through it.
Review weekly for the first month. Read the questions it couldn't answer, fix the gaps in your knowledge, and check every handover was picked up.

What it costs, worked through

All Klugon prices are in US dollars. The pricing page has the current plans.

For a clinic that mainly wants phone answering, the AI voice agent Starter plan is $99 a month (or $79 a month billed yearly) and includes 1,000 voice minutes, plus $9 a month for the phone number. The Business plan is $299 a month with 3,600 minutes and adds Google, Outlook and Apple calendar connections, Slack and Zapier, and bring-your-own-number over SIP. Extra minutes are $15 per started 100 minutes.

Here's a worked example. These are assumptions, not data about your clinic.

A small private GP clinic gets 25 calls a day, five days a week, plus 10 a day at weekends, which is roughly 145 calls a week, or about 630 a month.
Average call length with the AI is 2.5 minutes, so about 1,575 minutes a month.
That's above Starter's 1,000 minutes. On Starter, 575 extra minutes is six started blocks of 100 at $15, or $90. Total: $99 + $9 + $90 = $198 a month. On Business, it fits inside 3,600 minutes: $299 + $9 = $308 a month, with the calendar and Slack integrations included.

Now set that against the alternatives, again using assumptions you should replace with your own figures:

A part-time receptionist for evenings and Saturdays, say 15 hours a week. Whatever your local rate is, add employer's National Insurance, pension, holiday cover and training. The AI won't replace your daytime team, and shouldn't, but it can cover the hours you currently leave to voicemail.
A human answering service charged per call or per minute. Ask your provider for their rate and multiply by the after-hours calls you get. Many answering services take a message but can't book into your diary.

The useful question isn't "is it cheaper than a receptionist". It's "how many private bookings do we lose when nobody answers". Say you miss 15 calls a week out of hours, and one in four would have booked a £120 consultation. That's around £450 a week in bookings that went to another clinic. Your own numbers will differ; the missed calls guide shows how to estimate them.

Compliance for medical practices

None of what follows is legal advice. Talk to your data protection officer or adviser.

Health data is special category data

Under UK GDPR, health data is special category data, which needs a lawful basis and an additional condition for processing. The ICO's guidance notes that appointment details count as health data when they reveal something about a person's health. A booking for "menopause consultation" or "sexual health screen" does. So a phone or chat conversation with your clinic, and the booking it creates, will often contain special category data.

That has practical consequences:

Do a DPIA before going live, and record Klugon as a processor in your records of processing.
Update your privacy notice to say that calls and chats may be answered by an automated assistant, and who processes the data.
Collect the minimum. Tell the AI to take a name, phone number and the appointment type, not a description of symptoms. If a caller volunteers symptoms, the AI hands over rather than asking for more.
Decide retention. Conversations appear in your inbox. Delete what you don't need to keep, and export what belongs in the patient record into your clinical system yourself.

What we can say about Klugon's data handling

Klugon is operated by a German company under the GDPR. Our sub-processors, which include providers in the USA, are listed in the privacy policy, with transfers covered by standard contractual clauses. We don't claim the data is stored only in the EU. Klugon doesn't record calls. Calls are saved as conversations in your inbox, and on the default voice engine the saved log may mainly show the AI's side, so don't treat it as a full record of what the patient said. You can export conversations and records as CSV and delete individual records; full erasure of an account is done by email request.

Klugon isn't certified under ISO 27001 or SOC 2, and for US readers, it isn't a HIPAA-covered service and doesn't sign business associate agreements. A US practice handling protected health information should take that as a reason not to use it for patient calls.

The CQC and AI disclosure

Private clinics providing regulated activities in England register with the Care Quality Commission. The CQC's mythbuster on AI is written for GP services, but its points on governance, risk assessment and procurement are a sensible checklist for any registered clinic adopting an AI tool. Keep a short risk log of what could go wrong (a missed safety message, a wrong booking, a confidentiality slip) and what you did about each.

Tell patients they're talking to an automated assistant. Klugon's default greeting doesn't say so, which is why the setup checklist has you write your own.

How it compares with what clinics already use

Most private clinics already have a practice management system, often with an online booking page: Semble, Cliniko and WriteUpp are common in UK private practice. Those systems hold the clinical record and the diary. Klugon doesn't replace them and doesn't connect to them. What it adds is someone, or something, to answer the phone and the website chat when your staff can't, and to answer the admin questions that online booking pages don't.

Online booking page. Good for patients who already know what they want. It doesn't answer "do you take AXA" or "what do I bring to a taxi medical", and it doesn't pick up the phone.
Voicemail. Cheap and safe, but the private patient who gets voicemail on a Sunday night tends to book elsewhere.
A human answering service. A person on the line, which matters for distressed callers. Usually takes messages rather than booking.
An extra receptionist. Best for the complex calls and the patient relationship. The AI receptionist vs human receptionist comparison sets out the trade-offs honestly.

For many clinics the sensible pattern is the human team in the day, with the AI covering overflow, evenings and weekends, plus the website chat for the price and preparation questions people ask at 11pm.

How to measure whether it's working

Give it a month and look at numbers you can see, not impressions:

Calls and minutes on the analytics page, by hour and weekday. Are the out-of-hours calls being answered?
Bookings made by the assistant, and how many were later cancelled.
Share handed to a person. A clinic should expect a meaningful share of handovers, because results, medication and symptom calls all go to staff. A handover rate near zero would worry us more than a high one; it might mean the AI is answering things it shouldn't.
Questions without a good answer, listed on the insights page. Each one is a gap in your knowledge to fill.
Handovers left waiting. If staff aren't picking them up quickly, fix the notification routine before anything else.
Patient comments at the front desk. Ask a few patients how the call went.

Frequently asked questions

Can the AI triage patients or tell them whether they need a doctor?

No. It doesn't assess symptoms or urgency. When a caller describes symptoms or asks for medical advice, it reads your fixed safety message with 999 and NHS 111, and flags the conversation for your team.

Can it transfer a call to a receptionist or doctor?

No. Klugon has no live call transfer. It marks the conversation as needing a person, shows it in the inbox and notifies Slack or Zapier, and a member of staff calls the patient back.

Can an NHS GP surgery use it?

An NHS practice must check its GP contract, telephony framework, national reporting, DCB0160 and DTAC duties and its DSPT position with its ICB before using any AI phone tool. Klugon isn't an NHS framework supplier and doesn't have NHS assurance documents today, so it suits private practices better.

Does it work with EMIS, SystmOne or our practice management system?

No. It books into Klugon's own calendar. After setup on the Business tier it can check Google, Outlook or Apple calendars, but it doesn't connect to clinical systems.

Does it send appointment confirmations or reminders?

No. The patient gets a booking reference in the conversation itself. Klugon doesn't send confirmation emails, texts or reminders, so keep using your existing system for those.

Is it HIPAA compliant or covered by a BAA?

No. Klugon isn't HIPAA compliant, doesn't sign BAAs and holds no SOC 2 or ISO 27001 certification.

Does it record calls?

No. Calls appear as conversations in your inbox, but there's no audio recording, and the saved log may not include everything the caller said.

Will patients know they're talking to an AI?

Only if you tell them. The default greeting doesn't say it, so write a greeting that does.

Can we keep our existing practice number?

Yes, by forwarding calls from your number to a Klugon number through your phone provider, or over a SIP trunk on the Business tier. Numbers can't be ported, and a number can only be added after the free trial.

How long does the free trial last?

Seven days. A card is needed and is charged when the trial ends unless you cancel. During the trial you test the assistant in your browser.

If you run a private clinic and want to see how it handles your appointment types, start with your service list and safety message, then try the AI receptionist for small business in your browser. The clinics overview covers the wider picture, and the pages for dentists, therapists and vets show how the same rules play out elsewhere in healthcare.