Yes, a clinic can use an AI agent to answer booking requests after hours. The agent reads the clinic's WhatsApp line after the front desk goes home. It replies with the open slots and holds a slot in the booking calendar. Nothing is confirmed until the desk confirms it in the morning. A question for the doctor gets a short holding reply and waits for the doctor.
The 9pm booking request
Booking requests arrive on the clinic WhatsApp long after closing. A patient asks for a slot at 9pm. By the time someone replies at 9am, the patient has booked the clinic down the road.
The reply the patient needs is simple. Which slots are open, and can one be held. That is a lookup and a message, and an AI agent can do both.
What the agent does overnight
Our solutions page has an example chat for a clinic morning. The clinic manager asks Glimmer, the AI agent Glints builds for Singapore companies, what came in overnight.
In the example, 6 messages arrived after the 8:30pm close. Glimmer replied to the booking requests with the open slots and put holds in the calendar for 3 patients. One question was for the doctor. The patients are listed by initials. The manager confirms all three holds at 7:56am, before the doors open. See the example chat.
What waits for a person
- Bookings wait for the front desk. Glimmer holds a slot for the patient. The desk confirms it in the morning.
- Medical questions go to the doctor. A question such as an MC extension goes to the doctor. Glimmer tells the patient the clinic will reply.
- Claims wait for you. Glimmer fills in CHAS and Medisave claims. It does not submit a claim until you approve.
You set the rule once for each kind of message: "Ask first" or "Just do it". See how approvals work.
The paperwork after the last patient
The same agent can take the 9pm paperwork. In a second example, the clinic manager locks up and asks Glimmer to prepare the day's CHAS and Medisave claims and the Xero entries.
In the example, 31 patients visited that day: 12 CHAS visits, 3 Medisave claims under CDMP and 16 private. Glimmer pre-fills the claim entries from the day list and the receipts, six minutes after the ask. Nothing is submitted until the manager approves it. See the example chat.
The page puts the nightly claims and accounts at about 10 hours a week today. With Glimmer preparing them, it is about 1 hour 15 minutes. We count your own hours with you on the first call.
Patient data
Glimmer runs in a cloud account in the clinic's name, and you control who gets in. It reaches only the systems on your approved list, and every other connection is blocked and logged. Read where your data goes.
It works with your booking calendar, Google Sheets, your shared drive and Xero. For a system built for your clinic, we build the link with you.
Questions people also ask
Does the AI give medical advice?
No. A medical question goes to the doctor. The patient gets a short reply saying the clinic will respond.
Can it confirm a booking on its own?
Only if you set that rule. In the example, it holds the slot and the front desk confirms each one in the morning.
Does it reply in Mandarin or Malay?
Yes. A message in Mandarin or Malay gets a reply in the same language. See an example.
Is there a grant for this?
A ready-made AI chatbot from IMDA's pre-approved list is eligible for up to 50% support for an SME under EDGE, at most S$7,500. More details. A bigger build that joins up your systems can apply under EDGE's Large-scale Digital Transformation support. An example is your booking system, claims and accounts. One agent doing one task on its own is not eligible. EDGE supports up to 50% of eligible costs for an SME. More details. IMDA's AI x Digital Leaders also matches companies outside the tech industry with partner vendors. More details. Apply before you sign with any vendor. Read why.
Where to start
Ask your front desk how many new patients booked elsewhere last month because nobody replied overnight. A rough count is fine. Then book a call and bring it. We will tell you what Glimmer could answer on its own and what should wait for the desk.