Marketing reputation

Live chat on a clinic website: who answers it?

Staffed live chat, chatbots and enquiry forms compared for private clinic websites: staffing, safety, medical device limits and data protection.

Ava research teamPublished 6 min read

On this page
  1. The real question behind live chat
  2. Three options side by side
  3. Staffed chat: what it really takes
  4. Chatbots: scope is everything
  5. Be honest that it is software
  6. Designing the hand-over
  7. Do not underestimate a good form
  8. Rules for any chat channel
  9. How to choose
  10. Example: from chat bubble to booked consultation
  11. What to measure
  12. Frequently asked questions
  13. References

Key takeaways

  • Decide who answers before you add a chat window: a person, a chatbot, or a form that promises a reply.
  • Chatbots that assess symptoms or suggest treatment may be regulated as medical devices; keep a website bot to practical questions[1].
  • Patients are open to AI in healthcare but want it handled carefully, so be clear when they are talking to software[2][3].
  • Whatever answers, it must signpost urgent help, avoid clinical advice and collect as little health data as possible[4].

The real question behind live chat

A chat bubble in the corner of a clinic website looks like a small feature. In practice it is a promise: someone, or something, will answer quickly. If a visitor types a question at lunchtime and nobody replies until the next day, the chat window has made the clinic look slower than a simple form would have. So the decision is less about the technology and more about who answers, when, and what happens when nobody is available.

Websites are now a main route into healthcare. In the GP Patient Survey, the share of patients who said it was easy to contact their practice using the practice website rose from 47.9% in 2024 to 57.6% in 2026[5]. Private clinic visitors bring the same expectation of a quick, clear answer online.

GP patients finding it easy to contact their practice using the websiteEngland, GP Patient Survey
GP patients finding it easy to contact their practice using the websiteComparison. 2024: 47.9%; 2026: 57.6%.47.9%20242024: 47.9%57.6%20262026: 57.6%GP patients finding it easy to contact their practice using the websiteComparison. 2024: 47.9%; 2026: 57.6%.47.9%20242024: 47.9%57.6%20262026: 57.6%

Source: Ipsos (GP Patient Survey provider for NHS England), 2026[5]

Show the data
ItemValue
202447.9%
202657.6%

Three options side by side

Comparison of website enquiry options for a private clinic
OptionWhat it does wellWhere it strugglesSuits
Staffed live chatPersonal answers, can handle nuance and book on the spotNeeds cover during advertised hours; slow replies hurt trustClinics with a coordinator who can watch chat
Scripted chatbotInstant answers to fixed questions; predictable wordingFrustrating when questions fall outside the scriptClinics with many repeat practical questions
AI chatbotHandles varied wording; can answer from approved contentNeeds strict scope, testing and oversight; risk of wrong answersClinics willing to govern it carefully
Enquiry form with promised replySimple, structured, works out of hoursNo instant answer; depends on inbox disciplineSmaller clinics and complex enquiries

Staffed chat: what it really takes

Staffed chat works when it is treated as a front desk, not a side task. That means a named person on a rota during advertised chat hours, a target reply time, and the authority to book appointments or arrange call-backs. If the same person also answers the phone and greets patients, chat will be the first thing they drop at busy times.

  • Show chat only during staffed hours, and switch to a form with a reply promise when nobody is available.
  • Set a reply target, such as under two minutes, and check it weekly.
  • Give the chat team the same scripts as reception, including how to handle symptoms and urgent requests.
  • Make sure transcripts are stored securely and can be found if a patient complains.
  • Avoid chat on clinical pages if you cannot answer clinical-adjacent questions safely.

Chatbots: scope is everything

A chatbot is useful when most questions are practical and repetitive: opening hours, parking, whether you offer a treatment, how much a consultation costs, how to book. It becomes risky when it drifts into clinical territory. The MHRA says many software and AI products used in health and social care are regulated in the UK as medical devices[1]. A bot that asks about symptoms and suggests what might be wrong, or which treatment to have, may fall into that category, with obligations most clinics are not set up to meet.

  1. Write down the bot's job in one sentence, and the topics it must never answer.
  2. Build its answers from your approved wording: practical information, fees, services and booking.
  3. Make it say clearly that the visitor is talking to an automated assistant.
  4. Hand over to a person, or to a form, whenever the question is clinical, a complaint or anything the bot cannot answer.
  5. Signpost urgent help in fixed wording: NHS 111 for urgent advice, 999 in an emergency.
  6. Test it with fictional questions, including awkward ones, before launch and after every change.

AI Employees such as Ava follow this pattern on the phone: they answer practice information from approved wording, signpost urgent help and pass patient matters to reception without taking health details. Apply the same boundaries to any website bot.

Be honest that it is software

People are curious about AI but want it handled with care. 52% of UK adults expect AI to have a positive impact on healthcare[3], and 40% use AI chatbots for personal purposes at least monthly[6]. Yet healthcare is the area people most often say needs careful management of AI to prevent negative outcomes[2]. A bot that pretends to be a person, or that gives confident wrong answers, damages trust quickly. Label it, keep it in its lane and make it easy to reach a human.

52%

of UK adults expect AI to have a positive impact on healthcare[3]

Designing the hand-over

The weakest point of any chat set-up is the moment a conversation passes from the bot or the website to a person. If the visitor has to repeat everything, or the message lands in an inbox nobody checks, the clinic loses the enquiry. Design the hand-over as carefully as the answers.

  • Tell the visitor exactly what happens next: "A coordinator will reply by email by 10am tomorrow" is better than "Someone will be in touch".
  • Pass a short summary to the team: name, contact details, preferred contact method, the service mentioned and the question in the visitor's words.
  • Route hand-overs to one shared inbox or task list that is checked at fixed times, not to an individual's email.
  • Flag anything mentioning pain, bleeding, breathing or other urgent words for same-day review, while still showing the visitor the urgent care wording.
  • Close the loop: mark each hand-over as answered so nothing sits unresolved over a weekend.

Test the hand-over weekly by sending a fictional enquiry yourself and timing the reply.

Do not underestimate a good form

For many smaller clinics, a well-designed enquiry form beats chat. It works at 11pm, collects structured details, and lets the team reply with a considered answer. The weakness is speed, so make a specific promise ("We reply by 10am on the next working day") and keep it. Ask only for what you need to reply: name, contact details, preferred contact method, the service they are interested in and a short message. Add a line asking people not to include medical details.

Rules for any chat channel

  • Signpost urgent help on every chat screen and in every bot flow.
  • No diagnosis, no treatment advice, no interpretation of results.
  • Collect the minimum. Data concerning health is special category data and needs a specific legal condition[4].
  • Store transcripts securely, set a retention period and include chat in your privacy notice.
  • Sign a data processing agreement with the chat or bot supplier and check where data is stored.
  • Review a sample of conversations every month for accuracy and tone.

If you would not want a receptionist to say it on the phone, the chat tool should not say it either.

How to choose

  1. Count the questions

    For a month, log what people ask by phone, email and form. If most are practical and repetitive, a bot or FAQ page may help; if they vary, a person or a form is better.

  2. Check your cover

    Work out the hours someone can genuinely watch chat. If it is only an hour or two a day, use a form with a reply promise.

  3. Pick the smallest option

    Start with the simplest tool that solves the problem. You can add chat later.

  4. Set the boundaries

    Write the scope, the hand-over rules and the urgent wording before choosing a supplier.

  5. Pilot and measure

    Run it for six weeks and measure reply time, bookings, hand-overs and complaints.

Example: from chat bubble to booked consultation

Example: Harbour Skin Clinic (fictional) added staffed chat to its website, but the coordinator also ran reception and replies often took hours. The clinic changed to a scripted assistant for practical questions during the day, which handed anything else to an enquiry form with a next-morning reply promise, and kept staffed chat only on Saturday mornings when a coordinator was free. Questions about fees and parking were answered instantly, and the coordinator's replies were faster because the simple questions had gone.

What to measure

  • Median reply time for staffed chat and forms.
  • Share of bot conversations handed to a person, and why.
  • Enquiries that turn into booked appointments, by channel.
  • Complaints or corrections linked to chat answers.
  • Questions the bot could not answer, which tell you what to add to your website.

Frequently asked questions

Is an AI chatbot on a clinic website a medical device?

It depends on what it does. The MHRA says many software and AI products in health are regulated as medical devices[1]. A bot limited to practical information and booking is very different from one that assesses symptoms.

Do we have to tell visitors they are talking to a bot?

Yes, as good practice and for trust. Label it clearly and make it easy to reach a person.

Should live chat be available 24 hours a day?

Only if you can staff it or use a bot with a safe scope. Otherwise show chat during staffed hours and a form with a reply promise at other times.

Can chat collect medical history before a consultation?

It is better to use a secure pre-consultation form designed for health data, with a clear legal basis, than a general chat tool[4].

References

  1. GOV.UK (Medicines and Healthcare products Regulatory Agency) (2025). Software and artificial intelligence (AI) as a medical device (opens in a new tab) ↑

  2. GOV.UK (Department for Science, Innovation and Technology) (2024). Public attitudes to data and AI: Tracker survey (Wave 4) report (opens in a new tab) ↑

  3. GOV.UK (Department for Science, Innovation and Technology) (2024). Public attitudes to data and AI: Tracker survey (Wave 4) report (opens in a new tab) ↑

  4. legislation.gov.uk (The National Archives) (2026). UK GDPR Article 9: Processing of special categories of personal data (opens in a new tab) ↑

  5. Ipsos (GP Patient Survey provider for NHS England) (2026). 2026 GP Patient Survey results released (opens in a new tab) ↑

  6. GOV.UK (Department for Science, Innovation and Technology) (2024). Public attitudes to data and AI: Tracker survey (Wave 4) report (opens in a new tab) ↑

Book a demo

See Ava handle your calls.

Book a short demo. We’ll walk through your calls and workflow, and agree what Ava should take care of.

Book a demo