Guideazy
🏥
GUARDIAN for Healthcare — IndiaNOW ONBOARDING FOUNDING HOSPITALS

From the OPD Queue to Claim Settlement.
Every Record Linked. Every Decision Yours.

For nursing homes, hospitals, diagnostic labs and clinics — whether you run on a hospital software system or on paper and WhatsApp. GUARDIAN registers patients with their ABHA, turns the reports you already print into national-standard records, prepares pre-authorisations and claims, and follows up abnormal results — with your staff approving every step, and patients approving every share.

🏥 Nursing Homes & Small Hospitals 🏨 Multi-Specialty Hospitals 🧪 Diagnostic Labs & Centres 🩺 Clinics & Polyclinics ABDM · ABHA · NHCX · PM-JAY · Dr NTR Vaidya Seva · Aarogyasri
🇬🇧Working in UK healthcare? NHS and independent-sector providers work under different rules from India's, and we're building a UK version around them. Tell us what you need.
Register interest →
GUARDIAN's 4-Pillar Architecture
Four components working together, referenced by name throughout this page.
💬
Human-BOT
Talks to patients and staff on WhatsApp (a website chat is optional, never required), and routes every decision to the right person — the TPA desk, the doctor's desk, your data-protection lead.
⚙️
HAIYSAI
The workflow engine: builds national-standard records, prepares claims, flags abnormal results. It never diagnoses, and stops for a person wherever a decision matters.
🛡️
Data Privacy Shield
Reads documents inside its own boundary, then replaces patient names, phone numbers and ABHA numbers with tokens before any AI model sees them. Hosted in an Indian cloud region.
🔗
NEXUS
Connects to ABDM, NHCX, WhatsApp and your hospital software — and prepares complete claim packs for scheme portals that have no connection for outside software.
ℹ ABDM & Digital Health Incentives — What's Current
ABDM connects your hospital to patients' ABHA records and the national claims exchange. Its incentive scheme is short-term — so GUARDIAN is built to be worth it without one.
ABDM v3 APIs · DHIS Corrigendum 7 (April–September 2026) · Only KYC-linked records count
v3
The ABDM API version GUARDIAN is built on — required for incentives from July 2026
Sep '26
End of the current incentive terms, "subject to availability of funds"
100
Qualifying records a month deducted before any incentive is paid
📅 DPDP Act — When Each Part Applies to Hospitals
India's data protection law applies in stages. Hospitals handle more personal data than almost anyone — here's what each stage means.
Aug 2023 → Nov 2025 → Nov 2026 → May 2027
Nov '25
DPDP Rules notified; Data Protection Board set up
Nov '26
Consent Manager provisions apply
May '27
Core duties apply — consent, withdrawal, erasure, breach reporting
How It Works

See Exactly Where GUARDIAN Fits Into Your Hospital

Your software, printers and people stay as they are. GUARDIAN reads what you already produce, prepares records and claims, and nothing leaves the hospital until a named person approves it.

📥 What you already have
📱
Patients' ABHA app
Scan & Share at the counter
💬
WhatsApp
Patients, lab photos, staff
🖨️
Printed papers
Discharge summaries · bills · reports
🧪
Lab software or printer
HL7 results or printed reports
🏥
Hospital software (if any)
HL7 · webhook · API
No hospital software? Printing and WhatsApp are enough to start.
Capture
⟶
🛡Data Privacy Shield — hosted in India
GUARDIAN
NEXUS
Receives, reads and routes · ABDM · NHCX
Integration
↓
HAIYSAI
Builds records · prepares claims · flags results
Workflow
↓
Human-BOT
WhatsApp with patients and staff
Conversations
↓
✓ HITL Gate
Your staff approve · patients consent · UK patent-pending GB2615329.6
People decide
Approved
⟶
📤 Where it goes
🔗
ABDM — ABHA-linked records
Shared only with the patient's consent
📤
NHCX — insurers & TPAs
Pre-auth and claims, sent electronically
📦
Scheme claim packs
PM-JAY · Vaidya Seva · Aarogyasri — your desk uploads
💬
Patient follow-up
WhatsApp — never a diagnosis
📒
Consent & sharing ledger
Every share, per patient
Every output records who approved it, and when.
0
Raw patient names, phone or ABHA numbers seen by the AI model
0
Records, claims or shares released without a named person's approval
0
Diagnoses made by GUARDIAN — it prepares, flags and suggests; your doctors decide
Step 1 of 3

Which best describes your facility?

Choose yours. In Step 2, the problems that apply to you are marked ✓ — and Step 3 shows how GUARDIAN connects to what you already use.

✓
🏥
TYPE A — Nursing Home / Small Hospital
Paper files · Basic billing software · TPA desk
Roughly 20–60 beds. Registration, discharge summaries and claims are handled on paper, Word and a billing system. Insurance and scheme cases depend on one busy desk.
✓
🏨
TYPE B — Multi-Specialty Hospital
Hospital software (HIS) · Lab system · Many payers
A HIS and lab software hold the data, but claims, records and follow-ups still run on printouts. Often working towards NABH.
✓
🧪
TYPE C — Diagnostic Lab / Centre
Lab software or printed reports · Walk-ins and referrals
Produces reports all day. Needs them in patients' hands — and abnormal results followed up — without the counter ringing patients one by one.
✓
🩺
TYPE D — Clinic / Polyclinic
OPD only · WhatsApp · No IT team
A few doctors, busy OPD hours, prescriptions on paper, and appointments on WhatsApp. No servers and no IT person — and no wish for either.
Step 2 of 3 · Problems we solve

The problems GUARDIAN solves for hospitals, labs and clinics

Pick a problem to see what happens today, how GUARDIAN fixes it, where your staff stay in control — and what happens in the awkward cases.

More problems

Problem we solve · OPD registration

The morning OPD queue is typed in by hand — and almost no patient's ABHA is captured.

It's 9:30 on a Monday at Krishna Nursing Home in Guntur. Forty patients are waiting at one counter. Lakshmi, 54, spells her name for the receptionist, who types it — differently from her last visit — into the billing system. Nobody asks for her ABHA, so nothing done for her today can ever be linked to her national health record.

Sound familiar?

  • A long morning queue at a single counter
  • The same patient under three spellings, in three files
  • "Do you have your old file with you?"
  • Nothing ABHA-linked to show when a patient or insurer asks

Why it matters

  • The queue is the first thing patients judge you on.
  • Duplicate files split a patient's history — a doctor can miss an earlier result or an allergy.
  • Without a verified ABHA, a record can't be shared through ABDM or counted in any national digital-health scheme.
How GUARDIAN fixes it
1
📱
Lakshmi scans your counter's QR codeWith the ABHA app — or any ABDM-approved health app — on her own phone.
Patient
2
📥
Her verified details arriveName, age, gender and ABHA number — and a token number appears on her phone.
Automatic
3
🔎
GUARDIAN finds her fileMatched on her ABHA number. Possible duplicates are flagged, never merged automatically.
Automatic
4
🧑‍💼
Receptionist confirmsOne tap to confirm the match, or open a new file.
Human decision
5
🎫
Registered, ABHA capturedEverything done for her today can now be linked to her.
Done
🛡️

Where a person decides: the receptionist confirms every match. A patient without an ABHA can create one at the counter — only if she wants to — with an OTP on her Aadhaar-linked mobile.

How Scan & Share works — and what it doesn't do
  1. Your hospital is listed in the national Health Facility Registry (HFR). We help you register at setup if you aren't already.
  2. ABDM issues a QR code for each counter. You print it and put it on the desk.
  3. The patient scans it with her own phone and taps "share". Her verified profile reaches GUARDIAN and she gets a token number.
  4. No Aadhaar number is ever stored. For a new ABHA, the OTP is encrypted with ABDM's key and passed straight through.
  5. Scan & Share only registers the visit. Linking today's reports and summaries to her ABHA is a separate step — see "Records stuck on paper".

The current Digital Health Incentive Scheme pays nothing for Scan & Share on its own.

What happens when… — the situations we've planned for
  • No smartphone, or no ABHA — registered exactly as today; the ABHA can be added at a later visit
  • A child patient — the ABHA and every consent come from the parent or guardian
  • An emergency — treatment comes first; registration is completed afterwards by staff
  • A family sharing one phone — patients are never matched on phone number — the ABHA number is the key
  • An ABHA created with a mobile number only (not KYC-verified) — she is registered, but the file is flagged: records linked to it don't count for incentives
  • The same patient scans twice — she gets the same token; no second file is opened
  • Your billing system has no connection — GUARDIAN prints a registration slip with your file number for the receptionist
Less time at the counterVerified details arrive from the patient's phone instead of being typed.
One file per patientMatches are made on the ABHA number, not on spellings.
ABHA captured — with consentThe starting point for every linked record.

What setup involves: your hospital's HFR registration (we guide it) · ABDM approval for GUARDIAN at your facility · a QR standee at each counter · your receptionists try it with a few patients before go-live.

What we'd measure in your pilot: share of OPD patients with a verified ABHA, minutes per registration at peak hours, and duplicate files opened per week.

Works with the ABHA app and other ABDM-approved health apps patients already use, and your existing billing or hospital software.
See Scan & Share at your counter — book a demo →
For your IT team — how it connects
  • Scan & Share: ABDM calls GUARDIAN's registered callback (v3 /api/v3/hip/patient/share); GUARDIAN replies with the token number.
  • ABHA creation and verification use ABDM's M1 v3 APIs. Aadhaar numbers and OTPs are RSA-encrypted with ABDM's public key and never stored or logged.
  • Patient identifiers are replaced with tokens (FPE, NIST SP 800-38G) before any AI model sees them. Healthcare deployments are hosted in an Indian cloud region.
  • Going live needs ABDM's functional test, a security audit (WASA) and NHA approval — GUARDIAN completes these once; your facility is then linked.
Problem we solve · ABHA-linked records

Discharge summaries and lab reports are printed, filed and forgotten — none of them linked to the patient's ABHA.

Lakshmi's ultrasound report and discharge summary are printed on Thursday and put in her file. Six months later a specialist in Vijayawada asks what the scan showed. She doesn't have the paper, and the specialist can't see it.

Sound familiar?

  • Stacks of printed reports and summaries
  • Patients coming back for copies
  • "Digital records" means PDFs on one PC
  • Nothing to show when ABDM record sharing comes up

Why it matters

  • Patients lose continuity of care every time they see a new doctor.
  • You can't take part in ABDM record sharing — something patients and insurers increasingly expect.
  • Only records linked to a KYC-verified ABHA count when national incentive schemes run.
How GUARDIAN fixes it
1
🖨️
Your staff print as they do todayOr the lab photographs a report, or your hospital software sends it.
Your team
2
🛡️
GUARDIAN reads it privatelyText is read inside the privacy layer; names and ID numbers are replaced with tokens.
Automatic
3
📄
It builds the national-standard recordThe right type: discharge summary, lab report, prescription, OP note.
Automatic
4
🧑‍⚕️
Records staff check what's flaggedA smudged value or an uncertain patient match waits for a person.
Human decision
5
🔗
Linked to her ABHAShared only when she approves a request in her own app.
Done
🛡️

Where a person decides: records staff confirm flagged fields and the patient match before anything is linked — and the patient decides every share from her own ABHA app.

How record sharing works under ABDM
  1. Your records stay with you. Your hospital becomes a "Health Information Provider": GUARDIAN tells ABDM that a record exists for Lakshmi — not what's in it.
  2. Another doctor asks. The specialist's hospital requests her records through ABDM.
  3. Lakshmi decides. She sees the request in her ABHA app and chooses which record types and dates to share, and for how long.
  4. Only then is anything sent — encrypted end to end for that doctor alone.
  5. Corrections create a new version. A linked record is never silently changed.
What happens when… — the situations we've planned for
  • A handwritten prescription — stored as a scanned-document record and clearly labelled — not turned into typed text the doctor didn't write
  • A corrected lab report — the new version is linked and the old one marked as superseded
  • A patient without an ABHA — the record stays in your system and can be linked later
  • Sensitive results — HIV status, mental-health notes, pre-natal scans — handled under your hospital's confidentiality rules and never included in automated messages
  • A medico-legal case — records are kept and access restricted, as the law requires
  • The same report printed twice — recognised and linked once
No new typingRecords are made from what you already print.
Records follow the patientWith her consent, to any doctor on ABDM.
Ready for ABDMIn the national standard format, record type by record type.

What setup involves: your discharge-summary and report formats · the GUARDIAN Edge Agent on the PC your desk prints from (or WhatsApp photos, with no install) · ABDM approval for your facility · a test on past discharges before go-live.

What we'd measure in your pilot: records linked per month by type, share of patients with a KYC-verified ABHA, and staff minutes per discharge.

Works with whatever you print from today — Word, your billing software, your lab's printer — and HL7 messages if your hospital or lab software can send them.
See one of your discharge summaries become a linked record — book a demo →
For your IT team — how it connects
  • ABDM M2 (v3): care contexts linked with /api/hiecm/hip/v3/link/carecontext; consented requests answered with FHIR R4 bundles (NRCeS profiles), encrypted with ABDM's Fidelius scheme (ECDH Curve25519 + AES-GCM).
  • Text is read (OCR) inside the Data Privacy Shield. Clinical values stay readable; identifiers are tokenised.
  • The Edge Agent is a small signed installer: it captures print jobs, connects outbound only over TLS, and keeps no patient data after delivery.
Problem we solve · Pre-authorisation & claims

Your TPA desk retypes every case into insurer and scheme portals — then waits for the queries.

Lakshmi needs a laparoscopic cholecystectomy. If she's covered by a private policy through a TPA, your desk types her diagnosis, the procedure and the estimate into that TPA's portal. If she's covered by Dr NTR Vaidya Seva, the Vaidya Mithra types it all again into the Trust's system and uploads the scans. Two days later a query arrives: "ultrasound report not attached".

Sound familiar?

  • A different portal for every payer
  • Queries for a missing report
  • Patients waiting at discharge for approval
  • Weeks between discharge and payment

Why it matters

  • Every query delays cash — and a small hospital feels every day of it.
  • Retyping errors can mean a rejected or reduced claim.
  • The desk is the bottleneck at discharge — the moment patients remember.
How GUARDIAN fixes it
1
🖨️
The desk prints the papers as usualAdmission note, reports, discharge summary and final bill.
Your desk
2
🧾
GUARDIAN prepares the claimSuggests diagnosis and package codes, and checks the payer's list of required reports.
Automatic
3
🧑‍💼
Your TPA desk approvesCodes, amounts and attachments — nothing goes out before this.
Human decision
4
📤
Sent through NHCX — or packed for uploadPayer on NHCX: sent electronically. Scheme portal: a complete claim pack for your desk.
Automatic
5
📊
Every claim trackedStatus, queries and payments in one place.
Done
🛡️

Where a person decides: your TPA desk approves every pre-authorisation and claim. For Dr NTR Vaidya Seva, Aarogyasri and PM-JAY, your desk or Vaidya Mithra uploads the prepared pack through their own login — GUARDIAN never submits to those portals.

Two claim routes — and why GUARDIAN handles them differently
  1. NHCX is the national claims exchange, and many insurers and TPAs have joined it. For them, GUARDIAN sends your approved claim electronically — encrypted so only the payer can read it — and tracks the reply.
  2. State schemes and PM-JAY run their own portals. Dr NTR Vaidya Seva (Andhra Pradesh), Aarogyasri (Telangana) and PM-JAY's claims system have no route for outside software to submit claims today.
  3. For those, GUARDIAN prepares a claim pack: the fields in the portal's own order, the documents named and in sequence, and a checklist of what the package needs — so the upload takes minutes and nothing is missing.
  4. If a payer joins NHCX, the claim simply changes route. Nothing changes for your desk.
  5. GUARDIAN never stores portal passwords or OTPs.
What happens when… — the situations we've planned for
  • The payer sends a query — it goes to the desk with the original claim and documents attached
  • Partial approval — the difference is shown to the desk — and to the patient — before discharge, not after
  • Extension of stay or a change in procedure — a new pre-authorisation is prepared, so it isn't forgotten
  • The patient switches to paying herself — the claim is closed with the reason recorded
  • A package needs a specific report — flagged before approval, not after a rejection
  • A scheme portal is down — the pack waits; deadline reminders continue
  • A diagnosis or package code GUARDIAN isn't sure about — shown as a suggestion only — your desk or coder chooses
Fewer queriesMissing reports caught before submission.
One view of every claimNHCX and portal claims side by side, with status.
The desk approves, not retypesIts time goes to the cases that need judgement.

What setup involves: the list of payers you work with (and which are on NHCX) · your scheme empanelment details and package lists · NHCX registration for your hospital (we guide it) · a test on past claims before go-live.

What we'd measure in your pilot: minutes per pre-authorisation, queries per 100 claims, and days from discharge to settlement.

Works with NHCX for insurers and TPAs that have joined it, and prepared claim packs for PM-JAY, Dr NTR Vaidya Seva and Aarogyasri.
Walk through one of your recent claims with us — book a demo →
For your IT team — how it connects
  • NHCX uses the HCX protocol: /v1/preauth/submit and /v1/claim/submit, with the payer's answer arriving asynchronously on GUARDIAN's callback. Payloads are JWE-encrypted (RSA-OAEP, A256GCM) for the recipient.
  • Claims are FHIR ClaimBundles following the NRCeS profiles; the callback handler is idempotent on the correlation ID because NHCX retries.
  • For scheme portals no submission API is used — GUARDIAN produces the claim pack only. Diagnosis and package codes are suggestions until a named person approves them.
Problem we solve · Abnormal results

A patient's report comes back abnormal — and nobody notices she never came back.

Lakshmi's haemoglobin is 9.2 g/dL — flagged low. The printed report waits at the lab counter; she was told to collect it "any time". Nobody calls. Three months later she's back, more tired, with the same untreated anaemia.

Sound familiar?

  • Uncollected reports piling up at the counter
  • Front-desk staff calling patients from a register
  • Doctors unaware a flagged result was never seen
  • "Your report is ready" calls taking up the afternoon

Why it matters

  • A missed abnormal result is a clinical risk — and a hard conversation later.
  • A patient who doesn't return is a follow-up visit lost.
  • Routine calls use staff time that the front desk doesn't have.
How GUARDIAN fixes it
1
🧪
Your lab produces the reportPrinted, photographed, or sent by your lab software.
Your lab
2
🚩
GUARDIAN spots the flagged valuesUsing your lab's own reference ranges and flags — never its own.
Automatic
3
🧑‍⚕️
The doctor's desk reviewsDecides who needs follow-up, and how soon.
Human decision
4
💬
The patient gets a WhatsApp message"Please book a follow-up" — never the result or a diagnosis. Normal reports: "your report is ready".
Automatic
5
📅
Follow-up trackedNo reply? It goes back to staff to call.
Done
🛡️

Where a person decides: the doctor's desk decides every follow-up. Messages never contain results or diagnoses — they ask the patient to get in touch.

What happens when… — the situations we've planned for
  • A critical value — your lab's existing phone protocol stays in charge; GUARDIAN only adds a reminder that the call was made
  • A value GUARDIAN can't read clearly — never treated as normal — it goes to staff
  • The patient has opted out of WhatsApp — a call task is created for staff instead
  • Sensitive tests (e.g. HIV, pre-natal) — excluded from automated messages entirely
  • The patient prefers Telugu — messages go out in the wording your team has approved for each language
Fewer patients lostEvery flagged result has an owner.
No calling listsRoutine "ready" messages go out on their own.
Doctors see the flagsOn one worklist, not on paper.

What setup involves: your lab's report format(s) · the message wording you approve (English and Telugu) · your WhatsApp Business number · the list of tests that must never be messaged.

What we'd measure in your pilot: flagged results followed up within 14 days, staff calls saved, and uncollected reports at month end.

Works with printed or photographed reports, or HL7 results sent by your lab software.
See how a flagged result is followed up — book a demo →
For your IT team — how it connects
  • HL7 ORU^R01 results arrive through the Edge Agent (the OBX-8 field carries the lab's H/L flag), or values are read from printed reports inside the Data Privacy Shield.
  • Units are normalised (UCUM) and never re-scaled silently; reference ranges come from the report.
  • Patient messages use pre-approved WhatsApp templates; delivery is logged.
Problem we solve · Earlier records

Patients arrive without their earlier reports — so tests get repeated.

Lakshmi sees a physician at your hospital for the first time. Her last blood tests and scan were done in Guntur and Vijayawada; the papers are at home. To be safe, the doctor orders them again.

Sound familiar?

  • "Please get these tests done again"
  • Bags of papers from other hospitals
  • Patients refusing a repeat scan

Why it matters

  • Repeat tests cost the patient money and time.
  • Missing history can hide what matters.
  • "We found your last scan" is a service patients remember.
How GUARDIAN fixes it
1
📨
GUARDIAN asks for her recordsThrough ABDM — stating the purpose, record types and dates.
Automatic
2
📱
Lakshmi approves in her appNothing happens without her yes.
Patient
3
📥
Records arrive, encryptedFrom each hospital or lab that holds them.
Automatic
4
🗂️
A short summary for the doctorEarlier tests, results and dates — each linked to its source.
Automatic
5
🧑‍⚕️
The doctor decidesWhether a repeat test is really needed.
Human decision
🛡️

Where a person decides: the patient approves every request in her own app, and the doctor makes every clinical decision. The summary never offers a diagnosis.

What happens when… — the situations we've planned for
  • She says no, or ignores the request — nothing is fetched or stored
  • The consent period ends — fetched records are deleted automatically
  • A source hospital doesn't respond — the summary says so — no silent gaps
Fewer repeat testsWhen earlier results exist on ABDM.
History in one viewWith every item traceable to its source.

What setup involves: ABDM approval for GUARDIAN to request records for your facility · the purposes your doctors will use · a trial with consenting patients.

What we'd measure in your pilot: repeat tests avoided (doctor-reported) and consent requests approved versus sent.

Works with any hospital or lab that shares records through ABDM. Useful only where enough providers near you are live on ABDM.
Ask us how many providers near you share records today →
For your IT team — how it connects
  • ABDM M3 (v3): /api/hiecm/consent/v3/request/init, consent notifications, then /api/hiecm/data-flow/v3/health-information/request.
  • Received bundles are decrypted with GUARDIAN's key pair, identifiers tokenised, and deleted when the consent expires.
Problem we solve · NABH documentation

NABH certification stalls on hundreds of documents one quality manager has to write.

Krishna Nursing Home wants NABH Entry Level Certification to strengthen its insurer empanelments. The quality manager has a folder of old policies, photos of registers, and a consultant's quote.

Sound familiar?

  • Policies that don't reflect how the hospital really works
  • A consultant's invoice for templates
  • Accreditation plans that keep slipping

Why it matters

  • NABH helps with insurer empanelment and patient trust.
  • The documentation burden usually falls on one person.
How GUARDIAN fixes it
1
📂
Send what you haveForward policies by email or photograph registers on WhatsApp.
Quality manager
2
🔍
GUARDIAN checks against the standardsNABH Entry Level (2nd edition) — what's covered, what's missing.
Automatic
3
🧑‍💼
Quality manager decidesWhat to adapt and what to write.
Human decision
4
✍️
Drafts from your own factsPolicies and SOPs in your hospital's words.
Automatic
5
✅
Final sign-offEvery document approved before use.
Human decision
🛡️

Where a person decides: the quality manager decides what's needed and approves every document.

What happens when… — the situations we've planned for
  • The standards are revised — the gap check is re-run against the new edition
  • A register must be kept on paper — listed as a checklist item, not generated
A clear gap listAgainst the current standard.
First drafts, not blank pagesWritten from your hospital's facts.

What setup involves: your existing policies and registers · the scope you're applying for.

What we'd measure in your pilot: standards covered and weeks to readiness.

Works with documents by email, WhatsApp photos, or upload.
See a gap check on your own policies — book a demo →
For your IT team — how it connects
  • Standards kept as versioned configuration; drafts cite the source documents they were built from.
  • Contains little patient data; any that appears is tokenised.
Problem we solve · Hospital software

Your hospital software already holds the data — but it gets printed and read again.

Your HIS records every admission, result and charge. To get anything out, staff print it — and the vendor quotes months for a custom interface.

Sound familiar?

  • Printing data that already exists digitally
  • Vendor quotes for every new interface

Why it matters

  • Double handling of the same data.
  • Staff time on work software could do.
How GUARDIAN fixes it
1
🏥
Your HIS sends HL7 messagesAdmissions, results and charges — a setting your vendor turns on.
Your HIS
2
🔌
The Edge Agent receives themInside your network, sending out only.
Automatic
3
🔄
Mapped once, used everywhereRecords, claims and follow-ups start on their own.
Automatic
4
🧑‍💻
Your IT person approves the mappingTested on real messages before go-live.
Human decision
5
✅
No more printing to re-readWorkflows start from the HIS directly.
Done
🛡️

Where a person decides: your IT person or HIS vendor approves the message mapping, and every downstream record or claim still goes through its own approval.

What happens when… — the situations we've planned for
  • The HIS corrects or merges a patient — GUARDIAN updates the record — never duplicates it
  • The internet connection drops — messages wait securely on the agent and are sent when it's back; you're alerted if the queue grows
Automatic captureNo printing, no photos.
Your HIS unchangedOnly an export setting.

What setup involves: HL7 export enabled by your HIS vendor · a PC or small VM for the Edge Agent · a test with real messages.

What we'd measure in your pilot: messages processed without error and mapping exceptions per week.

Works with HIS and lab software that can send HL7 v2 (ADT, ORU, DFT), or a webhook/API.
Tell us which HIS you use →
For your IT team — how it connects
  • Mirth Connect MLLP listener on the Edge Agent; ACK/NAK per message; outbound TLS 1.3 (optionally IPsec VPN).
  • Per-hospital mapping profile; unknown codes queued for review.
Problem we solve · Digital health incentives

You don't know what the government's digital-health incentive should pay you — or whether it will continue.

Krishna Nursing Home heard that linking records to ABHA earns an incentive. Nobody can say how much of this month's work qualifies, or what should be paid.

Sound familiar?

  • Unsure which records qualify
  • Surprise at the payout — or none

Why it matters

  • The current terms are short-term. NHA's Digital Health Incentive Scheme (Corrigendum 7) covers April–September 2026, subject to funds.
  • The rules are specific: KYC-verified ABHAs only, per-patient caps, and a monthly baseline.
How GUARDIAN fixes it
1
🔢
Counts qualifying recordsBy category, after the scheme's caps.
Automatic
2
🧮
Calculates the expected amountUsing the scheme's own method.
Automatic
3
🧑‍💼
Accounts reviewAgainst the scheme's portal.
Human decision
4
📄
One monthly reportNo surprises.
Done
🛡️

Where a person decides: your accounts team checks every figure; claims on the incentive portal are made by your staff.

What happens when… — the situations we've planned for
  • The scheme isn't extended — the tracker is switched off; past reports are kept
ClarityWhat qualifies and why.

What setup involves: nothing extra — it runs on records GUARDIAN already links.

What we'd measure in your pilot: difference between expected and received.

Works with the ABDM records GUARDIAN links for you.
Ask us about the scheme's current status →
For your IT team — how it connects
  • Scheme rules are configuration, versioned per corrigendum.
Also included: secure video calls between doctors and patients when a conversation is needed · WhatsApp appointment reminders · a DPDP toolkit — consent and sharing ledger, withdrawal workflow, breach log and full audit trail · role-based access, so each staff member sees only what their role needs. On our roadmap: PM-JAY claims through NHCX as NHA opens that route more widely · automatic claim status from scheme portals if they offer a connection.
Step 3 of 3

How GUARDIAN connects to your facility

Type A — Nursing Home / Small Hospital

No Hospital Software Needed. Your Printer and WhatsApp Are Enough to Start.

GUARDIAN works from what your staff already produce. A small Edge Agent on the PC your desk prints from captures discharge papers and bills; WhatsApp brings in lab photos and patient messages; patients register by scanning your QR.

How NEXUS Connects a Nursing Home
📱
Patient's ABHA app
Scan & Share at the counter
→
ABDM callback
🖨️
Desk PC printer
Edge Agent · print capture
→
Read inside the Shield
🔒
Data Privacy Shield
Identifiers tokenised
→
After approval
🏛️
ABDM · NHCX · scheme portals
Linked · sent · or packed for upload
NEXUS: Scan & Share · Print Capture · WhatsApp → ABHA-Linked Records, NHCX Claims and Scheme Claim Packs.
The Edge Agent is a small signed program on one desk PC. It connects outwards only, keeps nothing after sending, and doesn't change how anyone prints. Claims to insurers and TPAs on NHCX are sent after your desk approves; for Dr NTR Vaidya Seva, Aarogyasri and PM-JAY, your desk or Vaidya Mithra uploads the prepared pack through their own login.
Problems GUARDIAN solves for you:
Workflow templates

Pick a Template. Connect It. Teach It Your Rules.

Every problem on this page comes as a designed workflow template for hospitals, labs and clinics, set up with you in a pilot. You do not start from a blank page: choose one, give it your name, connect your systems and upload your own rules. Your team approves every important step.

  1. ChoosePick a template for your problem
  2. RenameGive it your own name
  3. ConnectLink your systems through NEXUS
  4. Set the startA new record, a file, a message or a time
  5. TeachUpload your rules, letters and FAQs to its expert
  6. TestRun it on your past cases
  7. ShadowIt runs on real work; your team checks every step
  8. LiveSteps go automatic only once your team agrees they are right

Every template has its own expert: the SME Agent

Each workflow step can ask its SME Agent (subject-matter expert) a question, such as "which documents does our policy accept here?". It answers from your material only, and always shows where the answer came from.

If it is not sure, it says so, and the step waits for your team. A step moves on by itself only when the rules are met and the confidence is high enough. Important decisions always go to a person.

  • Current rulesRegulations and deadlines, kept up to date by Guideazy, with the date each rule starts
  • Your policiesYour SOPs, limits and ways of working
  • Your letters and formsThe wording and formats you already use
  • Your historyPast tickets, customers and orders, read from your systems with personal details hidden
  • Your team's decisionsWhat your approvers agreed or changed, so it learns your way
  • Nothing sharedEach customer's expert is separate. The AI works on tokenised data, not real names or numbers.
India

Front-Desk ABHA Registration (Scan & Share)

Receives the patient's verified profile when they scan the hospital's QR with their ABHA app, gives them a queue token, finds or creates their hospital file for the receptionist to confirm, and helps patients without an ABHA create one at the counter.

What it needs
Connects to
ABDM gateway (Scan & Share callback), ABHA service (M1), Hospital patient index (HIS or GUARDIAN's own)
Your team approves
Receptionist
Its expert learns from
Front-desk registration SOP, Patient ID and duplicate-file policy, Counter and department list with HFR details and 1 more
India

Clinical Documents to ABHA-Linked Records

Turns the discharge summaries, lab reports and prescriptions the hospital already produces into national-standard FHIR records, puts each on the records desk's worklist for approval, links approved records to the patient's ABHA, and sends them encrypted when the patient approves a request.

What it needs
Connects to
Edge Agent — print capture (Tier 1), Edge Agent — HL7 (Tier 2), WhatsApp / email / upload (Tier 0), Hospital patient index and 1 more
Your team approves
Medical records staff
Its expert learns from
Medical records SOP (what is filed, by whom, when), Sample of each document template the hospital prints, Department and doctor code list and 1 more
India

Pre-Authorisation & Claims (NHCX + Claim Pack)

Collects the case, checks the payer's document list, suggests codes and amounts for the TPA desk to approve, then submits through NHCX where the payer is on it — or prepares a complete claim pack for staff to upload to PM-JAY TMS, Dr NTR Vaidya Seva or Aarogyasri.

What it needs
Connects to
Edge Agent / upload, Patient coverage (desk entry or HIS), NHCX participant registry
Your team approves
TPA desk, Billing manager
Its expert learns from
Payer and TPA agreements (tariffs, packages, document lists), TPA desk SOP, Past query and rejection letters and 1 more
Where there is no API
GUARDIAN prepares the work; your team completes that step by hand, and it is recorded.
India

HIS Event Intake (HL7 Edge Agent)

Connects a hospital system that already sends HL7 v2 to GUARDIAN once, through the on-site Edge Agent, so admissions, discharges, results and charges start the other templates automatically — no printing or photos.

What it needs
Connects to
HIS (ADT)
Your team approves
Hospital IT / records lead
Its expert learns from
HIS vendor's HL7 interface specification, Code lists (departments, doctors, services, tests), Sample messages for each event type
India

Prior Records with Patient Consent (HIU)

When a doctor asks, requests the patient's earlier records from other providers through ABDM; only if the patient approves in their own app does GUARDIAN fetch them, list them for the doctor with links to each source, and delete them on the consent's expiry date.

What it needs
Connects to
Doctor / staff request, ABDM gateway (M3)
Your team approves
Requesting doctor, Patient
Its expert learns from
Hospital policy on requesting outside records, Consent request wording in patient languages
India

Abnormal Result Follow-Up

Puts every flagged lab result on the ordering doctor's worklist; when the doctor asks for a follow-up, GUARDIAN invites the patient back on WhatsApp — without mentioning results — and reminds staff if the patient has not booked.

What it needs
Connects to
LIS (HL7 ORU) or lab print / photo, Hospital patient index
Your team approves
Ordering doctor (or covering doctor)
Its expert learns from
Lab's critical-value phone protocol (for reference; unchanged), Follow-up booking SOP, Covering-doctor rota
India

NABH Entry-Level Documentation Assistant

Maps the policies and registers the hospital already has to NABH Entry Level standards, shows each gap with evidence, and drafts missing documents from the hospital's own facts for the quality manager to sign off.

What it needs
Connects to
Email / WhatsApp / upload, NABH standards (reference)
Your team approves
Quality manager, Medical superintendent
Its expert learns from
Existing policies and SOPs, Hospital facts sheet (beds, departments, services, staff roles), Document template / letterhead
India

Consent & Sharing Ledger

Records every place a patient's data goes and why, as a side effect of the other templates, and lets the hospital answer "who has my data?" for any patient in one view.

What it needs
Connects to
ABDM gateway
Your team approves
Designated data-protection person
Its expert learns from
Hospital privacy notice and consent forms, List of external recipients (labs, TPAs, insurers)
India

Patient Data Requests & One-Click Erasure with Legal Holds

Verifies the patient, lists the data held about them, applies the hospital's retention rules (medical records usually must be kept), and — after approval — erases what may be erased from GUARDIAN in one click by destroying the patient's personal key, keeping held records under a separate key until their retention ends.

What it needs
Connects to
Request channels
Your team approves
Designated data-protection person, Security lead (CISO or equivalent)
Its expert learns from
Hospital privacy notice and retention schedule, Data request / grievance SOP, Response letter templates
India

DHIS Incentive Tracker (only if DHIS continues)

Each month, counts ABHA-linked records that qualify under the DHIS rules in force, applies caps and the baseline, and gives accounts an expected amount to check against the DHIS portal. Switched off when no rule set covers the month.

What it needs
Connects to
Canonical Health Record, DHIS rule set
Your team approves
Accounts
Its expert learns from
DHIS portal statements received (for comparison)

10 templates. Each one is fully designed and documented for our engineering team. We set it up with you in a pilot, and it starts in shadow mode before any step runs on its own. Some steps, like final approvals, never run on their own.

Built for Cautious Hospitals

The Questions Hospital Owners Ask. Straight Answers.

Healthcare should be careful with new software. These are the concerns we hear most — and how GUARDIAN is designed around them.

🧰
Objection 1 — No IT
"We don't have hospital software or an IT person. This isn't for us."
▼
🧠
Objection 2 — AI and clinical risk
"AI shouldn't be anywhere near diagnoses or claims."
▼
🔐
Objection 3 — Patient data on the cloud
"Patient data can't leave India, and it can't go to a public AI."
▼
📉
Objection 4 — ABDM fatigue
"We heard ABDM would pay us. Nothing came of it."
▼
Platform Architecture

How the 4 Pillars Work Together — Healthcare

For your IT team — the four pillars, in the order data flows
📱
ABDM callbacks
Scan & Share · consent · record requests (v3)
🖨️
Edge Agent
Print capture · HL7 over MLLP · outbound-only
🏥
HIS / lab software
HL7 ADT · ORU · DFT · webhook / API
💬
WhatsApp · Email
Business API · monitored inbox
↓   Every connection authenticated · text read and identifiers tokenised before any AI model sees them   ↓
🔗
1 · NEXUS — Zero-Trust Integration Engine
ABDM M1–M3 on the v3 APIs (ABHA, record linking, consent-based sharing with Fidelius encryption); NHCX pre-auth and claims (HCX protocol, JWE-encrypted, asynchronous callbacks); WhatsApp Business API; HIS connectors; and claim packs for scheme portals with no submission API. Keeps the hospital-to-registry mapping (HFR facility and HPR doctor IDs).
🛡️
2 · Data Privacy Shield — Format-Preserving Encryption
OCR runs inside the Shield. Names, phone numbers, ABHA numbers and policy IDs are replaced with tokens (FF1, NIST SP 800-38G). Clinical values, codes, dates and amounts stay readable because the work needs them; they are protected by India-region private hosting, encryption and role-based access. Aadhaar numbers are never stored.
⚙️
3 · HAIYSAI — Agentic Workflow Engine
Builds FHIR R4 records by type (NRCeS profiles), prepares ClaimBundles and claim packs, flags abnormal results using the lab's own ranges, and keeps the consent ledger current. Suggests; never decides clinically.
💬
4 · Human-BOT — Conversations & the Multi-Consensus Gate
WhatsApp with patients and staff, and the approval gate where every record link, claim and share is signed off by a named person (UK patent-pending GB2615329.6).
⚙ Runs automatically
Receiving Scan & Share profiles and issuing tokens · Capturing prints, HL7 and photos · Reading documents (with confidence flags) · Building draft records and claims · Spotting lab-flagged values · "Report ready" messages · Logging every share · Tracking claim status on NHCX
👤 Always waits for a person
Patient-file matches · Linking a record to an ABHA · Every diagnosis or package code · Every pre-auth and claim · Uploads to scheme portals · Abnormal-result follow-up decisions · Withdrawal and erasure actions · Any record-sharing request (the patient decides in her app)
Clinical & Legal Safeguards — Human-in-the-Loop

GUARDIAN Doesn't Diagnose, Code or Submit on Its Own. Your Team Does.

In a hospital, the cost of a wrong automated decision is too high. So GUARDIAN is built to prepare and flag — and to stop for a person every time.

No clinical decisionsGUARDIAN never diagnoses, prescribes or decides treatment. Codes are suggestions; flags come from your lab's own ranges.
No results in messagesPatients are asked to get in touch — they are never sent a result or diagnosis by WhatsApp. Sensitive tests (e.g. HIV, pre-natal) are excluded from automated messages entirely.
Emergencies firstCare never waits for registration, consent capture or software. Staff complete the record afterwards.
Patients decide sharingEvery ABDM record request is approved by the patient in her own app — or by a parent or guardian for a child.
Records you must keep, keptMedico-legal cases and records other laws require are never erased on request; access is restricted instead.
1×
Each case keyed once — then reused for the record, the claim and the follow-up
0%
Raw patient identifiers seen by the AI model — FPE tokenisation
9
Hospital problems covered — each with a person at every decision
🇮🇳
Patient data hosted in India — on the cloud or in your own infrastructure
GUARDIAN Value Tool
Calculate What GUARDIAN Replaces at Your Facility
See which tools and manual hours GUARDIAN consolidates — in 2 minutes.
Open Value Tool →
Start Small

What a pilot looks like

You don't have to change your hospital at once. Start with one problem, prove it on real work, then decide.

1
Pick one problemThe one from Step 2 that costs you most — claims, registration, records or follow-ups.
2
Share what we needThe "What setup involves" line on that problem lists it — usually a few formats, a payer list and your consent wording.
3
Test on past casesYour staff try it on past discharges and claims before any patient sees anything.
4
MeasureTrack the numbers under "What we'd measure in your pilot" on that card.
5
DecideKeep it, extend it to another problem, or walk away.

Anything that touches ABDM or NHCX has a few extra steps

ABDM and NHCX connections follow NHA's standard certification — functional tests, security audit and approval. We manage this for you; founding hospitals go live on them as soon as it completes.

1 · Your facility is registeredIn the national Health Facility Registry (HFR), and your doctors in the professional registry (HPR) where needed.You, with our help
2 · GUARDIAN gets certifiedABDM's functional tests and an independent security audit (WASA) — done once by Guideazy, not by each hospital.Guideazy
3 · NHA approves go-liveProduction access is granted by the National Health Authority. We can't promise its timing.NHA
4 · Your facility is linkedYour counter QR codes and NHCX registration are set up, and the pilot starts.Together

Problems that don't touch ABDM or NHCX — abnormal-result follow-up, NABH documentation, the consent ledger — can start without these steps.

Less Typing at the Desk. Faster Claims. No Patient Left Without Follow-Up.

We're looking for our first founding hospitals. Start a 14-day free trial with follow-ups, the consent ledger, NABH documentation and claim packs — we'll add your ABDM and NHCX connections as part of onboarding.

Hosted in India. Built for the DPDP Act and ABDM's consent model. GUARDIAN never diagnoses, and nothing is submitted without a named person's approval.