On this page
Clinic receptionists can use AI to write the words they use every day: appointment messages, answers to common patient questions, phone scripts and Kiswahili versions of each. AI drafts the templates. The real patient details go in later, inside the clinic's own approved system. Patient records never go into a public AI tool, and a fluent person checks every translation before patients see it.
This guide is for receptionists, front desk officers and customer care staff at clinics and hospitals in Kenya. It is general information, not legal advice. Your facility's rules come first.
Why the front desk needs extra care
The front desk is where most patient data is first collected. Kenya's Data Protection Act, 2019 defines health data to include "data collected in the course of registration for, or provision of health services", and treats health status as sensitive personal data.
Section 29 of the Act says that before collecting personal data, a data controller should, as far as practicable, tell the person what is being collected, why, and who else may receive it. The Office of the Data Protection Commissioner (ODPC) uses a receptionist in its own example. Its December 2023 guidance note on health data describes a patient filling in a form at reception and says the receptionist should first explain what data is collected, why, how long it is kept, how it is protected and who it is shared with.
So the receptionist's words matter. AI can help you find clear ones.
Seven ways receptionists can use AI today
Each prompt uses the four Cs we teach at FuKazee: Clarity, Context, Constraints and Customisation. None uses a real patient's details.
1. Appointment reminder templates
Clarity: Write five short appointment message templates for a clinic: booking confirmed, reminder the day before, reschedule, missed appointment, and clinic closed early. Context: messages go by SMS from our clinic system to patients in Kenya. Constraints: under 160 characters each. Use only these placeholders: [NAME], [DATE], [TIME], [CLINIC PHONE]. Do not mention the doctor's specialty, the department or the reason for the visit. Customisation: warm and polite, in English and in simple Kiswahili.
Check: read each message as if someone else picked up the patient's phone. A reminder that names a clinic department can reveal a health condition. Your clinic manager approves the final set.
2. A patient FAQ from your clinic's own facts
Turn the facts below into a short FAQ for our waiting-room notice board and website. Questions to cover: opening hours, what to bring to a first visit, how to book or cancel, payment options, and what to do in an emergency. Use only these facts and add nothing. Where a fact is missing, write [CLINIC TO CONFIRM]. FACTS: [paste hours, address, booking number, accepted payment methods, insurance notes]
Check: every answer matches what your clinic actually does. Ask the manager to confirm anything about insurance or payment.
3. Phone scripts
Write a phone script for a clinic receptionist answering a new caller who wants to book. Include the greeting, the details to collect, how to explain the privacy notice in two sentences, and how to close the call. Add a separate box: "If the caller describes chest pain, heavy bleeding, difficulty breathing or any other emergency, stop the booking and [OUR EMERGENCY PROCEDURE]". Keep the script to one page.
Check: the emergency steps must come from your clinical lead, not from AI. Receptionists never assess how sick someone is.
4. Kiswahili and other languages
AI can draft Kiswahili versions of notices, scripts and FAQs. It can also make them sound stiff or change a meaning. Give the AI your approved English text and ask for natural, everyday Kiswahili, then have a fluent colleague check it word for word. For why AI is weaker in some languages, see who is building AI for African languages, including Kiswahili.
5. A plain-language privacy notice for the registration desk
Rewrite our privacy notice below as a short script a receptionist can read aloud in under one minute, then as a half-page handout. It must still say what we collect, why, how long we keep it, who we share it with, and who to contact with questions. Do not add or remove any promise. NOTICE: [paste your approved privacy notice]
Check: your data protection officer or manager approves the shortened version, because every promise in it must be true.
6. Replies to complaints and difficult moments
Write three calm replies a receptionist can use: to a patient upset about a long wait, to a caller demanding test results over the phone, and to a relative asking about another patient. Context: a private clinic in Kenya. Constraints: polite, short, no promises about clinical matters, and never share any patient information. Offer a next step each time.
Check: results and patient information go only through the route your clinic has approved. Our guide to training customer service teams on AI has more on tone.
7. Front desk admin
Notice-board signs, queue updates, a daily handover note for the evening shift (written from counts, not names), supply requests and short procedures are all good tasks. Keep each one free of patient details.
What receptionists should not do with AI
- No patient records in a public AI tool. That includes names, phone numbers, ID or insurance numbers, appointment lists, sign-in sheets and photos of forms. Removing the name is often not enough.
- No medical advice and no triage. If a patient asks what their symptom means, the answer comes from a clinician.
- No results or reports read out by AI tools. Results go through your approved process.
- No promotional messages to patient lists without proper consent. Section 37 of the Act says personal data may not be used for commercial purposes without express consent, unless a written law allows it and the person was told. The ODPC's health guidance says sensitive personal data must not be processed for direct marketing.
- No personal accounts for clinic work. Use only the tools and phones your clinic has approved.
- Report mistakes at once. If patient data goes somewhere it should not, tell your manager the same day. Under section 43 of the Act, a data controller must notify the Data Commissioner within 72 hours of becoming aware of a breach that carries a real risk of harm.
For the full picture for your facility, read how hospitals and clinics in Kenya can use AI safely, and Can staff use ChatGPT under Kenya's Data Protection Act?
Questions people ask
Can a receptionist use ChatGPT to reply to patients on WhatsApp?
Use AI to write reply templates with no patient details, then send replies from the clinic's approved account. Do not paste a patient's real message, name or number into a public AI tool to get a reply.
Can AI book appointments for us?
Some clinic systems include automated booking. Choosing one is a decision for your management and data protection officer, because it handles patient data. A free public chatbot is not a booking system.
What if a caller sounds very unwell?
Follow your clinic's emergency procedure and get a clinician at once. AI should never be used to judge how urgent a case is.
Is it okay to use AI for a Kiswahili version of our forms?
Yes, as a first draft. A fluent speaker must check it, and the clinic manager approves it, because a small change in wording can change what a patient agrees to.
Do receptionists need AI training?
In our view, yes. Receptionists collect much of a clinic's patient data at registration and send many of its patient messages. Short, practical training on templates and privacy helps them do both safely.
Sources
All sources last checked 15 November 2026.
- The Data Protection Act, No. 24 of 2019, ODPC copy: section 2 (health data, sensitive personal data), section 29 (duty to notify), section 37 (commercial use of data), section 43 (breach notification within 72 hours).
- Office of the Data Protection Commissioner, Guidance Note on the Processing of Health Data (December 2023): the receptionist example and direct marketing.
- FuKazee facts: AI training for healthcare in Kenya.
Where FuKazee fits in
FuKazee trains hospital, clinic and health NGO staff to use AI on the paperwork around care, and leaves every clinical decision with a qualified person. As our AI training for healthcare page explains, every exercise uses made-up patients or fully masked records, and names, ID and SHA numbers, phone numbers and dates of birth stay out. Short repeated sessions let one group cover while another trains. We have trained more than 3,000 people since 2022.
To talk about training for your front desk team, email lead@fukazee.com
Want guided support instead?
Not sure which programme fits? Tell us who you are and we will point you to the right path: corporate AI training, the ScholarShift Accelerator for postgraduates, or KCSE study plans and tuition.





