Starter pack · Hospitals & medical centres

A hospital where no call is dropped, entitlements do not stall, and leadership sees the picture every morning

Four AI agents for small and mid-sized private hospitals and multi-department medical centres. They take calls and LINE appointments for every department around the clock, prepare entitlement paperwork and first-pass insurance claims, follow up with patients after discharge, and summarise OPD queues, beds and complaints for leadership every morning — with every medical decision staying with the doctors and nurses. Pre-assembled, live in 2–3 weeks.

A fit if you…

  • Hospitals or medical centres of 50–200 beds with several OPD departments and a call center of 3–10 people, where lines back up and calls drop every morning
  • Referrals and insurance claims stalling because entitlement paperwork is incomplete, with staff chasing by phone and filling in the same forms twice
  • Leadership getting queue, bed and complaint reports days late, and PDPA worries about patient data keeping AI on hold

A day with this pack running

  1. 06:30The agent briefs the director and the head of nursing: today's OPD queues by department, free beds per ward, referrals waiting to be accepted, and yesterday's complaints still open
  2. 08:00Calls spike at the call center. The agent takes LINE and any call waiting over a minute, booking internal medicine and paediatrics, moving appointments, and explaining how to prepare. Anything urgent is transferred to a nurse at once
  3. 10:30For Monday's surgery patient, the agent checks whether the entitlement and insurance documents are complete, drafts the first-pass claim and lists what is still missing, then sends it to the entitlements desk to check before anyone contacts the patient
  4. 14:00A community hospital refers a patient in. The agent gathers the referral letter, the history and the entitlement details into one bundle, notifies the receiving department, and books the appointment as the doctor set it
  5. 17:00Twelve patients are discharged today. The agent sends the aftercare notes, the medication schedule and the follow-up date from the doctor's orders, by LINE or SMS on whichever channel the patient consented to
  6. 20:00It asks patients discharged 3 days ago how they are, using the question set the nurses approved. Normal answers close the round; an unusual answer, or medication not taken, reaches the nurse on duty with the history at once

The output you get

Queues and record summaries (draft)Waiting for the doctor
Patients waitingOPD 42 people · 18 min average wait
Draft record summaries12
Not yet reviewed3

The AI drafts only from notes that already exist · a doctor reviews and signs every one

An illustrative example — the names and figures are invented so you can see the shape of the work; the real thing is built around your business.

Work arrivesThe agent briefs the director and the head of nursing: today's OPD queues by department, free beds per ward, referrals waiting to be accepted, and yesterday's complaints still open06:30
The AI agent does itCall center & LINE appointments+ 3 more agents in the pack
A person decidesYou approve in one clickUrgent symptoms, diagnosis, and appointments a doctor has to assess first are transferred to a nurse or the call center every time.
Result-60%dropped calls and long waits

Example: Hospitals & medical centres

LIVE DEMO

1 · Work arrives

Calls spike at the call center. The agent takes LINE and any call waiting over a minute, booking internal medicine and paediatrics, moving appointments, and explaining how to prepare. Anything urgent is transferred to a nurse at once

2 · The AI agent does it

  • Call center & LINE appointments
  • Entitlements and first-pass claims
  • Discharge follow-up & reminders
  • Morning report for leadership

3 · You just approve

Urgent symptoms, diagnosis, and appointments a doctor has to assess first are transferred to a nurse or the call center every time.

Approve in one click

Result

-60%

dropped calls and long waits

What is in the pack

4 AI Agent already assembled

Answers calls and LINE around the clock: booking, moving and confirming appointments, and saying which department a given problem is seen in, what the clinic hours are and how to prepare, from the doctors' schedules and a knowledge base the hospital approved. When the lines are busy, the agent takes the calls that have waited longest before a person does.

Urgent symptoms, diagnosis, and appointments a doctor has to assess first are transferred to a nurse or the call center every time.

Before the appointment or the admission it checks whether the entitlement documents, the insurance card and the referral letter are complete, drafts the first-pass claim on each insurer's own form, and tells the patient what else to bring — and gathers referral cases from the sending hospital into a single bundle.

The entitlements desk checks and confirms every bundle before it goes to the insurer. The AI does not decide whether cover applies.

Sends the aftercare notes, the medication schedule and the follow-up date from the doctor's orders after discharge, reminds patients to take their medication on schedule, and asks how they are at 1, 3 and 7 days with the question set the nurses approved. An unusual answer reaches the nurse on duty with the history at once.

The AI does not change medication, does not recommend treatment, and does not answer symptom questions itself. Every unusual case has to be answered by a nurse or a doctor.

Every morning it pulls from the HIS, the queue system and the complaint channels and puts one page together: OPD queues by department, average wait, bed occupancy, referrals waiting to be accepted, and complaints still open, flagging whatever is off the usual.

The report is input to a decision. Staffing, beds and how a complaint is answered stay with leadership and the department heads.

All included

What is in the box

  • Conversation templates for booking, moving appointments and routing to the right department, per OPD department, with the rules on what the AI must not answer
  • Templates for checking entitlement documents and drafting first-pass claims on the forms of the main insurers the hospital accepts, extendable by you
  • Discharge messages, medication reminder schedules and follow-up question sets, split by condition group and approved by the nurses
  • An approval screen for nurses and the entitlements desk: unusual symptom cases, claim bundles waiting to be checked, and calls the AI passed to a person
  • A PDPA data map, consent forms per channel, a retention and deletion policy, and a per-patient access log
  • The morning report and a weekly dashboard (calls, queues, claims, follow-ups, complaints), and 3 training sessions for the call center, the nurses and the entitlements desk

Go live

Live in 2–3 weeks

Step 1 / 3 · Collect each department's rules and set up PDPA

Week 1

  1. 1Week 1

    Collect each department's rules and set up PDPA

    Three to four hours with leadership, the head of nursing, the call center and the entitlements desk to collect the doctors' schedules, the routing between departments, the insurer forms, the follow-up rounds, and the line the AI must not cross, plus the data map and the consent forms.

  2. 2Week 2

    Connect the HIS and queue system, test on real calls

  3. 3Week 3

    Go live department by department and tune

-60%

dropped calls and long waits

from the hospitals we have done, within the first month, because the AI takes LINE and the morning overflow before a person does

-50%

time to prepare a claim per case

from checking documents ahead and drafting the claim so staff only check and confirm; usually visible in month 2

-25%

readmissions within 30 days

from medication reminders and follow-up after discharge for every patient; usually visible in month 2–3, depending on the condition group

Price and scope

One price, paid once, to install the whole pack, plus a small monthly fee for models, telephony and keeping the system running. All in, about the same as 1–2 call center staff per month. Fixed scope as set out in the pack, with no hidden costs.

Not included

  • Building or migrating a HIS/EMR, and full real-time integration with insurers or the NHSO
  • A Voice Agent taking every call in place of the whole call center (this pack mainly takes overflow calls and LINE, and can be extended in month 2–3)
  • Medical image analysis, diagnostic assistance, and the doctors' own medical records work
Get a quote for this pack

Where you go when you grow

Once you run several hospitals in a group, or need deep integration with the HIS and insurers, move up to Agentic Workflow Design and Implement & Coach built around your hospital, or add the Customer service on every channel and Leadership sees it first bundles, without starting over.

Safety in this business

  • The AI does not diagnose, does not change medication, and makes no medical decision in any case. Every symptom question goes to a nurse or a doctor, urgent cases are transferred to a person immediately rather than waiting for a round, and whether cover applies is decided by staff and the insurer.
  • Health data is sensitive data under PDPA. We build a data map, take consent per channel before contacting anyone, set how long data is kept and how it is deleted, and log which patient's data was accessed by whom and when, so it can be audited.
  • Patient data is not sent out to train models. The model sees only the data that task needs, and you can use a model that runs in Thailand or a Private LLM inside the hospital when the data governance committee requires it.

FAQ for Hospitals & medical centres

How far can the AI go in suggesting a department, or discussing symptoms?

It can only say which department the hospital routes a given problem to, along the routing the nurses approved. It does not diagnose and does not recommend treatment. If a patient describes something that counts as urgent, the system transfers them to a nurse straight away, with a summary of what was said.

Will thousands of patient records be safe, and PDPA-compliant?

Health data is sensitive data. This pack comes with a data map, consent forms per channel, a retention and deletion policy, and per-patient access logging. The data is not used to train outside models, and you can use a Private LLM inside the hospital.

We already run a HIS and a queue system. Do we have to change them?

No. We connect to the HIS and the queue system where they offer an API or can export data. If parts of it cannot connect, we run the pack's appointment schedule and reports alongside them, and the team still sees everything in one place.

Does the AI decide insurance claims in place of our staff?

No. The AI only checks whether the documents are complete and drafts the claim on the form. The entitlements desk has to check and confirm every bundle before it is sent. Whether cover applies is still decided by staff and the insurer.

What if a patient wants a real person, or the AI gets it wrong?

A patient can ask for a person at any time, by voice or in writing, and the system transfers them to the call center with the conversation summarised. Every AI answer is logged, and there is a weekly report of the questions it could not answer, so your team can fix the knowledge base itself straight away.

How do we start, and what do we need to prepare?

Start with a free 30-minute call, no obligation, to see whether this pack fits your hospital. Then have the doctors' schedules, the routing between departments, the main insurer forms, and 1–2 months of recorded calls or chats ready. We handle the rest within 2–3 weeks.

A free 30-minute call to see if the hospital pack fits you

Tell us how many departments you have, what HIS you run, when calls drop, and why claims stall. We will say straight out whether this pack is enough or whether it should be designed around you. No obligation.

Want to see it working first? Open the sample workspace for a business like yours — this exact team of agents is running in it. Sample data, no sign-up.

Free consultation

Book a free call 30 minutes