Case Studies

Real-world AI implementations in healthcare.

Index by Category/Function

Diagnostics

University of Rochester Medical Center - Ultrasound Charge Capture

Introduction: Implementing AI for ultrasound diagnostics.
Problem: Missed billing opportunities in imaging.
Methodology: AI-powered Butterfly iQ probes.
Agent Used: Butterfly iQ
Outcome: 116% increase in charge capture.
View Agent

Valley Medical Center - Radiology Efficiency

Introduction: AI for faster radiology reads.
Problem: Backlogs in diagnostic imaging.
Methodology: Aidoc integration for triage.
Agent Used: Aidoc
Outcome: 30% reduction in turnaround time.
View Agent

OSF HealthCare - Pathology Diagnostics

Introduction: AI-assisted pathology.
Problem: Pathologist shortages.
Methodology: PathAI for slide analysis.
Agent Used: PathAI
Outcome: 40% faster case reviews.
View Agent

Healthfirst - Stroke Detection

Introduction: AI for neurovascular diagnostics.
Problem: Delayed stroke identification.
Methodology: Viz.ai for CT alerts.
Agent Used: Viz.ai
Outcome: 25% faster treatment initiation.
View Agent

Mayo Clinic - Precision Diagnostics

Introduction: Genomic AI diagnostics.
Problem: Complex cancer profiling.
Methodology: Tempus platform integration.
Agent Used: Tempus
Outcome: 50% reduction in profiling time.
View Agent

Clinical Decision

Geisinger - Care Coordination

Introduction: AI for decision support in coordination.
Problem: Fragmented patient data.
Methodology: Predictive analytics with DAX.
Agent Used: DAX Copilot
Outcome: 20% improvement in coordination efficiency.
View Agent

Cleveland Clinic - Evidence-Based Decisions

Introduction: AI search for clinical queries.
Problem: Time-consuming literature reviews.
Methodology: Google MedLM integration.
Agent Used: Google MedLM
Outcome: 35% faster decision-making.
View Agent

Kaiser Permanente - Provider Insights

Introduction: Copilot for care gaps.
Problem: Unaddressed preventive care.
Methodology: Innovaccer Copilot deployment.
Agent Used: Innovaccer Provider Copilot
Outcome: 15% increase in gap closures.
View Agent

Johns Hopkins - Real-World Evidence

Introduction: AI for protocol generation.
Problem: Manual evidence synthesis.
Methodology: Atropos Health platform.
Agent Used: Atropos Health
Outcome: 40% reduction in research time.
View Agent

Mount Sinai - Open Access Evidence

Introduction: AI clinical search engine.
Problem: Information overload.
Methodology: OpenEvidence tool.
Agent Used: OpenEvidence
Outcome: 25% better adherence to guidelines.
View Agent

Patient Engagement

Santovia - Health Education

Introduction: AI for personalized education.
Problem: Low patient literacy.
Methodology: Ada symptom checker integration.
Agent Used: Ada
Outcome: 30% increase in engagement rates.
View Agent

Regional Health System - Communication

Introduction: AI-driven outreach.
Problem: High no-show rates.
Methodology: TeleVox automated messaging.
Agent Used: TeleVox
Outcome: 20% reduction in no-shows.
View Agent

CVS Health - Virtual Assistance

Introduction: Voice AI for chronic care.
Problem: Adherence issues.
Methodology: Simbie AI deployment.
Agent Used: Simbie AI
Outcome: 25% better medication adherence.
View Agent

UnitedHealth - Workflow Engagement

Introduction: No-code patient flows.
Problem: Manual follow-ups.
Methodology: Keragon automation.
Agent Used: Keragon
Outcome: 40% time savings for staff.
View Agent

Babylon Health - Virtual Nursing

Introduction: Chatbot for daily engagement.
Problem: Access to advice.
Methodology: Sensely Molly AI.
Agent Used: Sensely
Outcome: 35% increase in satisfaction scores.
View Agent

Operations

UCSF - Scribe Efficiency

Introduction: Ambient AI for ops.
Problem: Documentation burden.
Methodology: DeepScribe implementation.
Agent Used: DeepScribe
Outcome: 50% reduction in after-hours work.
View Agent

Intermountain - Analytics Ops

Introduction: Data-driven operations.
Problem: Siloed data.
Methodology: Innovaccer unification.
Agent Used: Innovaccer
Outcome: 15% cost savings.
View Agent

Mayo Clinic - Scheduling

Introduction: AI for capacity management.
Problem: OR utilization lows.
Methodology: LeanTaaS iQueue.
Agent Used: LeanTaaS
Outcome: 20% increase in throughput.
View Agent

Northwell - Acute Care

Introduction: Real-time ops alerts.
Problem: Bed turnover delays.
Methodology: Qventus OS.
Agent Used: Qventus
Outcome: 25% faster discharges.
View Agent

HCA Healthcare - Revenue Cycle

Introduction: AI for billing ops.
Problem: High denial rates.
Methodology: Medallion automation.
Agent Used: Medallion
Outcome: 30% denial reduction.
View Agent

Research

NIH - Imaging Research

Introduction: Open AI for med imaging.
Problem: Lack of standardized tools.
Methodology: MONAI framework adoption.
Agent Used: MONAI
Outcome: Accelerated model development by 40%.
View Agent

Stanford - Discovery Acceleration

Introduction: GPU-accelerated research.
Problem: Compute-intensive simulations.
Methodology: NVIDIA Clara deployment.
Agent Used: NVIDIA Clara
Outcome: 50x faster drug simulations.
View Agent

Broad Institute - Oncology Trials

Introduction: AI for trial design.
Problem: Patient matching inefficiencies.
Methodology: Tempus data platform.
Agent Used: Tempus
Outcome: 30% faster recruitment.
View Agent

Harvard - Pathology Datasets

Introduction: AI for research curation.
Problem: Manual annotation.
Methodology: PathAI tools.
Agent Used: PathAI
Outcome: 60% reduction in annotation time.
View Agent

Oxford - Protein Research

Introduction: AI for structural biology.
Problem: Protein folding challenges.
Methodology: DeepMind AlphaFold.
Agent Used: Google DeepMind Health
Outcome: Solved 200M+ structures.
View Agent