Hospital AI and operations glossary
The vocabulary hospital software procurement runs on — accreditation, clinical systems, revenue cycle and AI — defined in one place, without marketing language.
- NABH accreditation(National Accreditation Board for Hospitals & Healthcare Providers)
- India's national hospital accreditation programme, run by NABH under the Quality Council of India. It assesses a hospital against standards covering patient care, medication management, infection control, human resource management and continuous quality improvement, and requires documented evidence — including current, verifiable staff training records — for every standard. The 6th Edition is the current standard set.
- JCI accreditation(Joint Commission International)
- The international counterpart to NABH, widely pursued by hospitals in the GCC and by Indian hospitals serving international patients. JCI is generally regarded as more prescriptive on documentation and measurement, which makes continuously computed quality KPIs materially easier to maintain than annually reconstructed ones.
- HIS(Hospital Information System / HMS)
- The core transactional system of a hospital: registration, appointments, admissions, clinical records, orders, pharmacy, billing and discharge. A HIS is the system of record. Layers such as CRM, LMS, analytics and claim automation sit on top of it and consume its events rather than replacing it.
- EMR / EHR(Electronic Medical Record / Electronic Health Record)
- The patient's clinical record. An EMR is the record held by one organisation; an EHR is the record shared across organisations. AI scribing tools write into the EMR/EHR — a note that stays inside a separate app has not been documented.
- AI medical scribe(ambient clinical documentation)
- Software that listens to a consultation, transcribes it, and generates a structured clinical note for the clinician to review and sign. An ambient scribe captures the natural doctor-patient conversation; a dictation scribe captures a doctor's spoken summary. Accountability for the final note always remains with the clinician.
- Healthcare LMS(learning management system for hospitals)
- A platform that assigns, delivers, assesses and evidences mandatory staff training — infection control, BLS, patient safety, department SOPs — for every employee, with reporting that accreditation bodies accept. Its defining requirement is not course delivery but proof: who completed what, when, and how their understanding was verified.
- AI proctoring
- Automated invigilation of an online assessment using webcam face detection, audio monitoring, tab-switch detection and copy/paste prevention. It is what turns a training quiz score from an honour-system number into defensible accreditation evidence.
- HRMS sync
- A bidirectional connection between the HR system and the training platform, so a new joiner is enrolled automatically the moment they appear in HRMS and courses are assigned by role, department and location. Without it, training assignment silently drifts out of date as staff join, move and leave.
- Patient journey(care journey)
- A structured, milestone-based programme a patient moves through over weeks or months — antenatal care, chronic disease management, pre- and post-surgical pathways. Journeys are the healthcare equivalent of a sales pipeline, except milestones are clinical, sequenced by date, and can legitimately pause.
- Healthcare CRM
- A customer relationship platform modelled on patients and care journeys rather than deals and stages, capturing enquiries from website, chatbot, call centre, walk-ins and QR codes, communicating over WhatsApp, and integrating with the HIS. See the comparison with generic CRM for the criteria that separate them.
- TPA(Third Party Administrator)
- An intermediary that processes cashless insurance claims between a hospital and an insurer, operating its own portal, document formats and turnaround times. Most Indian hospital revenue-cycle delay accumulates inside TPA portals rather than at the insurer.
- Accounts receivable days(AR days)
- The average number of days between raising a claim or bill and receiving payment. Indian hospitals commonly sit at 60-90 AR days on insurance business; every day of AR is working capital trapped outside the hospital.
- Claim denial and rejection
- A rejection is a claim returned without adjudication, usually for documentation or coding defects; a denial is an adjudicated refusal to pay. Both are cheaper to prevent with pre-filing validation than to appeal after the fact.
- ICD-10 / ICD-11(International Classification of Diseases)
- The WHO coding systems used to classify diagnoses on clinical records and claims. Correct coding is simultaneously a clinical documentation task and a revenue task: under-coding loses legitimate revenue, mis-coding causes rejections.
- Revenue cycle management(RCM)
- Everything between a patient episode and the money landing: documentation, coding, billing, claim filing, payer follow-up, denial handling and collection. In hospitals the binding constraint is usually documentation speed and quality, not collection effort.
- Text-to-SQL
- A technique where a language model converts a plain-language question into a database query, allowing a hospital leader to ask 'what was OT utilisation last month by specialty?' without a BI analyst. Safety depends on scoping the generated query to a single tenant's data and a governed metric catalogue.
- UCDM(Universal Clinical Data Model)
- A normalised schema that hospital data from different HIS vendors and spreadsheet exports is mapped into, so KPIs mean the same thing across sites. It is what makes group-level comparison possible when each hospital's source system is different.
- DPDP Act(Digital Personal Data Protection Act, 2023 (India))
- India's data protection law governing the processing of digital personal data, including patient data. Practical obligations for hospitals include purpose limitation, consent, data-minimisation, breach notification and the ability to honour access and erasure requests — which is why tenant isolation, role-scoped access and audit trails are procurement requirements rather than nice-to-haves.
- Zero data retention
- A contractual and technical configuration in which an AI provider does not store the inputs or outputs of a request beyond the duration of processing, and does not use them for model training. For clinical AI this is the difference between an acceptable and an unacceptable processing chain.
- Whitelabelling
- Delivering software under the customer's own brand. In hospital software this ranges from a branded subdomain, logo, colours and certificates (usually standard) to a custom root domain, branded transactional email and a brand-owned app-store listing (usually a scoped, paid add-on).
- Multi-tenancy and tenant isolation
- An architecture where one software instance serves many hospitals while each hospital's data is strictly partitioned. Tenant isolation is the guarantee that no query, export or misconfigured permission can surface another hospital's patients.

