Platform Overview

Four Platform Layers. One Pipeline. Data Becomes Action.

DataCloud processes the mess. NovaHub serves it in real time to any application. Application Engine gives it structure. AI Platform makes it act. We built all four because no one else would -- and bolting vendors together doesn't work.

The pipeline IS the product

Watch clinical data flow from raw hospital systems to actionable patient engagement

01
HIS Raw Data
150+ hospital formats
02
Clean & Structure
ELT, MDM, dedup, fact build
03
Clinical Intelligence
Flags, scores, care gaps
04
NovaHub API
Real-time patient 360
05
Patient Engagement
WhatsApp, SMS, Email, Voice
06
Measurable ROI
5.3x average return
0+
Hospital formats ingested without custom code
0+
Care gap definitions mapped to clinical guidelines
<0h
Hospital go-live -- config, not a project
0.0x
ROI -- the platform pays for itself in under a week
Real-Time
Patient 360 APIs powering CRM, HIS, and AI agents
Any Source
Connect to existing databases, data lakes, or HIS directly

The Complete Architecture

Seven layers, each independently valuable. But the real advantage is vertical integration -- data flows through the entire stack without translation layers or sync jobs.

1. Data Sources
HIS
Hospital Information Systems
EMR
Electronic Medical Records
ERP
Financial & Operations
Lab Systems
LIMS & Pathology
Claims
Insurance & TPA
2. DataCloud -- Data Engineering
Ingestion
150+ formats
ELT
Clean, dedup, structure
MDM
Golden records
Dims & Facts
Star schema build
Metrics
Flags & scores
Cohorts
Patient segments
Care Gaps
200+ protocol definitions
Journeys
Patient timelines
3. PrismML -- ML Intelligence Layer (Add-on)
Experimentation
Multi-framework model selection
Training + Inference
DataCloud-native scoring
Model Governance
Promotion gates, drift, rollback
Explainability
SHAP + calibration + fairness
Enriched Scores
Risk, engagement, financial
4. NovaHub -- Real-Time API Gateway
Arrow Flight
Columnar streaming
Storage Engine
Purpose-built reads
Identity Resolution
Phone/ID lookup
Filtering
Multi-dimension
Atomic Updates
Zero downtime
5. Application Engine -- Schema-Driven Backend
Entity Schemas
Metadata-driven
Workflows
Automation engine
RBAC + ABAC
Role & attribute access
Preset System
Deploy CRM verticals
Multi-tenant
Hospital isolation
6. PEP + CRM Suite -- Engagement & Operations
PEP
Campaigns & ROI
CRM Suite
Patient/Doctor/B2B/Sales
Multi-Channel
WhatsApp, SMS, Email, Voice
Workflows
Clinical automation
Reports
ROI tracking
7. AI Platform -- Runtime & Agent Framework
Unified Execution
Same pipeline, every mode
Policy Engine
LLM proposes, policy decides
Trust & Safety
Injection, toxicity, scope
Agent Builder
No-code agent creation
Any LLM
Pluggable backend

The Pipeline IS the Product

Follow a single patient record from raw hospital export to booked follow-up appointment. Every step below is automated.

1

Ingest

150+ formats

Raw data arrives in 150+ formats -- CSV dumps, HL7 messages, proprietary exports. Schema auto-detection figures out the structure. No custom code per hospital.

2

Clean & Structure

ELT + MDM

Raw data gets cleaned, renamed, deduplicated, and structured into dimensions and fact tables. MDM resolves patient identities across systems -- 3 fragmented records become 1 golden patient. The ELT layer does the heavy lifting so downstream queries are fast.

3

Enrich

200+ care gaps

Now the data gets smart. Clinical flags, risk scores, care gap detection against 200+ protocol definitions. Cohort builder segments patients. Journey assembler stitches timelines together.

4

Serve

Real-Time APIs

NovaHub delivers curated patient data to any application -- CRM, HIS, EMR, AI agents. Phone number to complete clinical profile, in real time, via config-driven APIs.

5

Engage

33% conversion

PEP generates personalized follow-up messages and runs care gap campaigns. CRM Suite handles tickets, leads, and agent workspace. WhatsApp, SMS, Email, Voice -- the patient's preferred channel.

6

Measure

5.3x ROI

Every message, every call, every appointment -- attributed back to revenue. You know exactly which campaign, which channel, which care gap generated the booking.

Platform Layers. Deep Dive.

Each layer works on its own. But the compound effect of vertical integration is why competitors can't just copy one piece.

DC

DataCloud

The Data Engineering Layer

The hard problem nobody wants to solve. 150+ hospital data formats, each with its own quirks. DataCloud ingests them all, resolves patient identities into golden records, computes clinical flags, builds cohorts, detects care gaps, and assembles patient journeys. All config-driven -- no custom code per hospital.

Explore DataCloud
Key Capabilities
  • Schema auto-detection from raw hospital exports
  • ELT: clean, rename, deduplicate, type-cast, drop columns
  • MDM: 3 records become 1 golden patient via exact-match-first
  • Dimension & fact table build: star schema, surrogate keys
  • Metrics engine: clinical flags, risk scores, pre-computed aggregations
  • Cohort builder + care gap detection: 200+ protocol definitions
NH

NovaHub

Real-Time Data for Every Application

Your CRM, HIS, EMR, and AI agents all need curated patient data in real time. NovaHub delivers it. Register a dataset in DataCloud, get an API endpoint automatically. No custom code, no integration projects. Curated, resolved, intelligence-ready data for any application.

Explore NovaHub
Key Capabilities
  • Arrow Flight: columnar data streaming at wire speed
  • Purpose-built storage engine for high-throughput clinical data
  • Identity resolution: phone number to unified patient record
  • Multi-dimension filtering across clinical datasets
  • Atomic updates: swap data versions with zero downtime
  • Config-driven: register a dataset, get an API -- no code changes
AE

Application Engine

The Schema-Driven Backend

We got tired of writing CRUD. So we made it so that defining an entity schema generates the API endpoints, validation rules, UI components, and AI tools automatically. Every hospital gets the same engine with schema-per-tenant isolation. No forks.

Explore Application Engine
Key Capabilities
  • Schema-driven: entity definitions generate everything
  • Workflow engine: visual builder for clinical automation
  • Event system: triggers on create, update, delete, schedule
  • Multi-tenant: complete data isolation per hospital
  • AI agent integration: tools auto-generated from schemas
  • Preset system: deploy entire CRMs from config
MA

PEP

Patient Engagement Platform

Marketing automation that actually works in healthcare. Cohort-based segmentation, multi-channel journeys, care gap campaigns, and full-funnel ROI attribution -- from patient identification to booked appointment to revenue.

Explore PEP
Key Capabilities
  • NL-powered cohort builder with clinical dimensions
  • Visual journey builder with branching and scheduling
  • 4-channel delivery: WhatsApp, SMS, Email, Voice AI
  • 200+ care gap definitions with automated closure campaigns
  • Morning intelligence brief: prioritized daily action plan
  • Full-funnel ROI attribution: per-campaign, per-channel, per-condition
CR

CRM Suite

Operational CRM for Healthcare

CRM presets for any healthcare vertical -- Patient, Doctor, B2B, Sales, or build your own. Deployed from config, not custom code. Tickets, leads, tasks, SLAs, telephony, AI agents, Patient 360 workspace. One codebase. Different schemas per hospital.

Explore CRM Suite
Key Capabilities
  • Preset-driven deployment: any CRM vertical from configuration
  • Specialist AI agents with coordinator routing
  • Metadata-driven UI: auto-generated forms, tables, views
  • Tickets, leads, tasks with SLA tracking and escalation
  • Telephony integration: click-to-call, screen pop, call outcomes
  • Knowledge bank with AI agent assist for live call support
AI

AI Platform

Runtime Kernel & Agent Framework

Every AI interaction -- chat, voice, WhatsApp, autonomous background jobs -- runs through the same runtime with the same safety guardrails. Pluggable LLM backend, auto-generated tools, trust and safety enforcement. The LLM proposes, the policy layer decides, the executor enforces.

Explore AI Platform
Key Capabilities
  • Unified execution across all deployment modes
  • Policy engine: LLM proposes, policy decides, executor enforces
  • Trust & safety: injection protection, toxicity filtering, topic boundaries
  • Pluggable LLM: swap models without changing agents
  • Agent builder: create and deploy agents without code
  • Intent routing, conversation memory, cost budgets
ML

PrismML

ML Intelligence Layer — DataCloud Add-on

Hospital-localized ML training and batch inference on DataCloud data. Models are trained on your patient population, scored nightly, and written back to the data platform — enriching every patient profile, care gap, and campaign with ML intelligence. Governed by promotion gates, drift monitoring, and fairness analysis.

Explore PrismML
Key Capabilities
  • Experimentation: multi-framework model selection and tuning
  • Training + inference: DataCloud-native, nightly population scoring
  • Promotion gates: 3-gate governance before production deployment
  • Drift monitoring: feature/score drift with auto-retraining
  • Explainability + fairness: SHAP, calibration, demographic parity
  • Closed loop: ML scores enrich Patient360, cohorts, and campaigns

Why We Built It This Way

Not another healthcare dashboard. Not a horizontal CRM with a hospital skin. We made specific architectural bets -- and they paid off.

CD

Configuration, Not Customization

Every hospital gets the same platform. Differences live in config -- schemas, workflows, agents, reports. No fork. No custom branch. We refused to build a services company disguised as a product company.

DF

Show the Data Flowing, Not the Screens

Dashboards don't treat patients. We show data flowing -- from hospital systems through clinical intelligence to patient follow-ups. If you can't trace the pipeline, you're just looking at screenshots.

AI

AI-Native, Not AI-Washed

AI is in the architecture, not bolted on for the pitch deck. Agent framework, toolset registry, typed state actions, strict validation. Every AI interaction is testable and auditable. That matters when you're dealing with patient data.

HW

Healthcare-Specific, Not Horizontal

200+ care gap definitions. Clinical flags. Hospital presets. Hindi/English multilingual. Safety guardrails that actually stop the AI from diagnosing. Built for Indian healthcare from day one -- not a generic SaaS with a hospital skin.

Hospital Onboarding: 48 Hours, Not 48 Weeks

Traditional healthcare CRM vendors quote 3-6 months per hospital. We quote 48 hours. Here is how.

01

Connect

Hour 0-4

Schema auto-detection reads the hospital's data exports. Formats identified, fields mapped, pipeline config generated. Most columns auto-detected.

Auto-DetectField Mapping
02

Process

Hour 4-12

MDM resolves patient identities. Clinical flags computed. Care gaps detected. Cohorts built. Journeys assembled.

MDMCare GapsCohorts
03

Configure

Hour 12-24

PEP and CRM Suite configured with hospital-specific entities, workflows, and AI agents. Multi-channel communication set up.

PEPCRMWorkflows
04

Live

Hour 24-48

Morning brief lands in staff inboxes. First AI-drafted patient messages go out. ROI tracking starts. The hospital is live and seeing value.

Morning BriefROI

Numbers. With Context.

Data Engineering

150+
Hospital formats -- each one took work to map
3:1
Dedup ratio -- 3 fragmented records become 1 golden patient
200+
Care gap definitions mapped to clinical guidelines
18mo
Clinical data depth per patient

Data Delivery

Real-Time
Curated data delivered to CRM, HIS, EMR, and AI agents
Zero Code
Register a dataset, get an API -- no custom integrations
Any Source
Connect to existing databases, data lakes, or HIS directly
Zero Downtime
Atomic updates -- old data serves until new data is ready

Patient Outcomes

33%
Outreach-to-appointment -- not opens, actual bookings
40%
No-show reduction with Voice AI reminders
5.3x
Platform ROI -- follow-up revenue vs. platform cost
<6 days
ROI payback -- the platform earns its keep in a week

See It Running. 25 Minutes. No Slides.

7 demos. Real hospital data flowing through real infrastructure to real patient engagement and measured ROI.