Real-Time Clinical Intelligence
130M+
ER Visits / Year
36.4M
Hospital Admissions
14.7
Readmitted in 30 Days
$52.4B
Wasted Annually
Source: AHRQ HCUP, CMS, CDC NCHS
The Status Quo
Raw HL7 ADT Message
MSH|^~\&|HOSP_ADT|JEFFERSON|RECV||202603122014||ADT^A03|MSG001|P|2.3
PID|1||MRN4721^^^JUH||GARCIA^MARIA||19580415|F|||
PV1|1|I|4EAST^4E12^01||||ATT^PATEL^R|||MED||||ADM|||
DG1|1||I50.9^Heart failure, unspecified^ICD10|||ADRaw HL7 message
Hospital sends a wall of coded segments — PID|1||MRN123^^^HOSP||SMITH^JOHN — that no human can read at a glance.
Maybe it arrives
If the hospital knows who the PCP is. If the systems are connected. If the interface is built. Most of the time, it doesn't.
Buried in an inbox
If it arrives, it lands alongside hundreds of other messages. No prioritization. No clinical context. No suggested action.
No one connects the dots
The PCP's office staff has no idea this CHF patient was discharged with a new medication list that conflicts with what they prescribed last week.
The Solution
A single clinical event triggers an autonomous pipeline that notifies every stakeholder and generates a compounding data asset.
ER Event
ADT Admission
Atticus AI Pipeline
Providers
Fax / Direct / SMART on FHIR
Payers
Real-time dashboard
Employer Groups
TPA integration
Patients & Families
App + SMS alerts
Pharma & Research
Trial recruitment
De-Identified Data Asset
Every event compounds into a licensable dataset for pharma, research, and payer analytics.
ER Event Detected
ADT Admission Message
AI Pipeline
Providers
Fax / Direct / SMART on FHIR
Payers
Real-time dashboard
Employer Groups
TPA integration
Patients & Families
App + SMS alerts
Pharma & Research
Trial recruitment
De-Identified Data Asset
Compounds into licensable dataset
AI Agent Architecture
Each agent in the pipeline operates autonomously — detecting events, resolving identity, assembling records, synthesizing alerts, delivering to stakeholders, and managing ongoing care. No human in the loop for routine operations.
Monitors real-time ADT feeds from Health Gorilla, Particle Health, and regional HIEs. Identifies clinically significant events — ER admissions, discharges, transfers — and triggers the pipeline.
Processes 320M+ patient events nationwide
Probabilistic patient matching across systems. Resolves the patient's identity using demographic, clinical, and insurance data. Confidence-tiered: prefers no alert over a wrong-patient alert.
Multi-source matching with safety-first design
Queries multiple data sources to assemble the patient's complete clinical picture — medications, labs, problem list, vitals, prior encounters. Normalizes everything to FHIR R4.
Aggregates from EHRs, pharmacies, labs, claims
Reads the assembled clinical record and generates role-specific, condition-specific alerts. A PCP gets medication context. A cardiologist gets procedure history. A payer gets utilization data.
$0.01-0.03 per synthesis via Claude
Routes alerts to every stakeholder through the channel they use. Fax and Direct message for providers. Dashboard for payers. Push notifications and SMS for patients and families.
Multi-channel: fax, Direct, push, SMS, dashboard
When a patient receives an SMS alert and downloads the app, this agent auto-enrolls them into a condition-specific care program. Monitors connected devices, manages medication reminders, coordinates with the AI Health Assistant.
Autonomous programs for all 15 conditions
320M+
Patients accessible via data networks
42M
Covered lives in founder network
$0.01-0.03
AI synthesis cost per event
15
Clinical conditions at launch
The Product
Every alert is role-specific, condition-specific, and actionable.
AI Clinical Summary
Maria presents with acute CHF exacerbation. BNP elevated at 1,840 pg/mL (baseline 320). Weight gain of 8 lbs over 2 weeks suggests fluid retention. Current medications include Lisinopril 20mg, Metoprolol 50mg, Furosemide 40mg. Last cardiology visit was 4 months ago. Consider medication titration and follow-up within 7 days of discharge.
This alert was generated by Atticus AI and reviewed for clinical accuracy. Source data from Health Gorilla ADT feed.
AI-synthesized clinical summaries delivered via fax and Direct message. Free for every provider. Always.
Role-specific: PCPs see medication context, specialists see procedure history
Condition-specific: CHF alerts differ from COPD alerts
Source-referenced: every data point linked to its origin
Delivered in the format providers already use (fax, Direct)
Why Now
Why hasn't someone done this already? Because the prerequisites didn't exist until now.
Impossible 2 years ago
LLMs can read clinical records and generate role-specific, condition-specific alerts for $0.01-0.03 per event. This was science fiction in 2022.
Didn't exist 5 years ago
Health Gorilla and Particle Health created real-time ADT feeds covering 320M+ patients. For the first time, a single company can see clinical events nationwide.
CMS penalties increasing
CMS readmission penalties are rising. TEFCA is enabling nationwide data exchange. The regulatory environment is pushing — not blocking — innovation.
The Insight
When a patient gets an SMS alert, they download the app and auto-enroll into an autonomous AI care management program. 10-50x cheaper than human-coach models.
ADT Event
ER Visit Detected
SMS Alert
Sent to Patient
App Download
Patient Enrolls
Care Program
Auto-Enrollment
AI Assistant
24/7 Support
Cost Per Member Per Month
$0.17-0.50
Atticus AI
$15-50
Human Coaches
30-100x cost reduction with autonomous AI programs
Autonomous Care Includes
Connected devices (500+ via HealthKit/Health Connect)
AI Health Assistant with full clinical context
Instacart: condition-specific nutrition plans
GoodRx: automated medication savings
Clinical Coverage
Every major driver of readmission and cost.
Adult (10) · Pediatric (3) · Maternal (1) · Equity (1)
| # | Condition | Category | Admissions |
|---|---|---|---|
| 1 | Septicemia / Sepsis | Adult | 2,220,000 |
| 2 | Heart Failure (CHF) | Adult | 1,140,000 |
| 3 | Osteoarthritis | Adult | 1,010,000 |
| 4 | Pneumonia | Adult | 741,000 |
| 5 | Diabetes w/ Complications | Adult | 700,000 |
| 6 | Acute Myocardial Infarction | Adult | 658,000 |
| 7 | Cardiac Dysrhythmias | Adult | 620,000 |
| 8 | Cerebral Infarction (Stroke) | Adult | 533,000 |
| 9 | COPD | Adult | 500,000 |
| 10 | Acute Renal Failure | Adult | 400,000 |
| 11 | RSV / Bronchiolitis | Pediatric | 75,000 |
| 12 | Preterm Birth / NICU | Pediatric | 370,000 |
| 13 | Pediatric Asthma | Pediatric | 96,000 |
| 14 | Severe Maternal Morbidity | Maternal | 55,000 |
| 15 | Sickle Cell Disease | Equity | 115,000 |
Source: AHRQ HCUP NIS, CDC NCHS, CMS HRRP, March of Dimes
Pediatric Care Gap
Payer care management programs are designed to call the patient directly. For pediatric conditions, the patient is a child. The program structurally cannot reach the people who make care decisions — parents, pediatricians, and school nurses.
The standard model fails here
Pediatric asthma, NICU discharge, RSV — these conditions represent significant readmission volume and cost. But no one has solved care coordination for this population because the patient can't participate in their own care management.
Dr. Nicole Green, CMO & Co-Founder
25 years in pediatric emergency medicine. The clinical experience to design care programs specifically for pediatric populations — an underserved segment where payer demand exists but no adequate solution is available.
Atticus Routes Alerts to Care Decision-Makers
Parents / Guardians
Primary care decision-makers for pediatric patients
Pediatrician
Clinical oversight and medication management
School Nurse
Daytime monitoring and emergency response
Family Circle
Extended caregivers with shared visibility
Competitive Position
No major care management platform has solved pediatric routing. This is a structural gap in the market — not a feature gap. Atticus addresses it from day one with condition-specific protocols for pediatric asthma, NICU discharge, and RSV, with alerts routed to the adults who manage the child's care.
Market Size
Every hospital admission in America is an event Atticus can process. Every payer is a potential customer.
36.4M admissions x PMPM across all payers + employer market + data asset
$18.2B
300M+ commercially insured lives x weighted PMPM tiers
$5.4B
Initial payer contracts, ~2M covered lives in first markets
$120M
320M+
Patients in data networks
42M
Lives in founder network
$14.2B
Digital health funding (2025)
54%
Captured by AI companies
Business Model
Provider alerting costs pennies to deliver. Making it free is the beachhead — every provider who receives an alert becomes a node in the network. Revenue comes from payers, employers, pharma, and API access.
Primary revenue
PMPM contracts across Alerting, Care Management, and Premium tiers. Beachhead: regional payers, then national.
Self-insured employers
Third-party administrators managing self-insured employer populations. Same PMPM model, different buyer.
De-identified data
Longitudinal care transition data for clinical trial recruitment, real-world evidence studies, and population health research.
Year 3+
Third-party developers and health systems building on the Atticus intelligence layer via API access.
Payer Pricing Tiers
PMPM
Real-time ADT event alerts
AI-synthesized clinical summaries
Fax + Direct message delivery
Provider portal access
PMPM
Everything in Alerting
Autonomous care programs (15 conditions)
Patient app + family circle
AI Health Assistant
Connected device monitoring
PMPM
Everything in Care Management
Instacart nutrition integration
GoodRx medication savings
Payer engagement dashboard
De-identified data analytics
Custom condition protocols
Cost Per Event
$0.15-0.30
At scale (Year 3+)
Gross Margin
86-93%
At scale
Payer ROI
$6-15
Saved per $1 spent
Independent Revenue Stream
Every event Atticus processes generates de-identified, longitudinal care transition data. This dataset compounds over time and becomes independently valuable — a second business inside the first.
Physically separated from PHI
Separate databases, separate access controls, separate networks. Persistent pseudonymous tokenization — not just logical separation. Protected via BAA clause with all data partners.
The Flatiron model, applied to care transitions
Flatiron built a $1.9B business on curated oncology data. Tempus reached $6.1B+ with AI + clinical data licensing. Atticus builds the equivalent for care transitions — across all 15 conditions, not just one.
Revenue Timeline
Year 1
Accumulate
Year 2
License
Year 3+
Scale
Real-world evidence from care transitions. Patient cohort identification for trial recruitment. Post-market surveillance data.
Longitudinal care transition patterns across conditions, geographies, and demographics. Readmission risk modeling at population scale.
Benchmarking care transition outcomes across networks. Identifying high-cost utilization patterns before they compound.
Team
Built the last generation of care management software. Now building the AI-native replacement.
CEO & Founder
Built and ran care management at Medecision, serving 42 million covered lives. Deep relationships across the health plan executive landscape with direct access to the decision-makers who buy care management solutions.
CMO & Co-Founder
25 years in emergency medicine. Brings firsthand understanding of what happens at the point of care — the discharge gaps, the missing context, the follow-up failures. The clinical credibility to design care programs and validate AI outputs.
Competitive Moat
Each moat reinforces the others. The longer Atticus operates, the harder it becomes to replicate.
Not a legacy platform with AI bolted on. Every decision, every architecture choice is AI-first. This compounds.
Free alerting is the beachhead. Every provider who receives an alert becomes a node in the network. The network effect compounds — more providers means more data means better AI means more providers.
Consumer viral loop. Patient invites family. Family members become patients when they need care.
Longitudinal care transition data compounds over time. Year 2+: pharma, life sciences, population health research.
Market Comparables
$14.2B in digital health funding in 2025 — highest since 2022. AI companies captured 54% of all funding. AI-native startups command 83% premium on round sizes.
Recent AI Healthcare Valuations (2024-2025)
AI clinical documentation. Doubled valuation in 4 months. 150+ health systems. June 2025.
AI + precision medicine + clinical data. IPO. Dual revenue: SaaS + data licensing. June 2024.
AI healthcare infrastructure. Acquiring companies. $105M ARR, 150% YoY growth. 2025.
AI-native MSK care. Premium multiple for AI-native approach. AI-native commands premium. 2024.
Google spinoff. AI drug discovery. AI-first from day one. March 2025.
Core Comparable Companies
| Company | Valuation |
|---|---|
Collective Medical 2020 | $650M |
Bamboo Health 2021 | ~$1.5B |
Flatiron Health 2018 | $1.9B |
Innovaccer 2021 | $3.2B |
Tempus AI 2024 | $9B+ |
PMPM
Recurring SaaS
Free model
Network Effects
Year 2+
Data Asset
30-50x EV/ARR
AI-Native
AI companies command 30-50x EV/ARR vs 10-15x for non-AI health tech (2-5x premium)
Ask Atticus
Ask anything about the strategy, technology, market, or business model.
Ask anything about Atticus