Applications of AI Agents in Medicine and Healthcare

Transforming Medical Education, Clinical Practice, and Patient Care

28th August 2025, UHS Rohtak

Ashish Makani�KCDH-A, Ashoka University

AI in medicine has very old roots !

Introduction to AI in Healthcare

Artificial Intelligence is transforming healthcare across multiple domains:

Clinical Decision Support:

AI systems that analyze patient data to assist in diagnosis and treatment planning

Medical Education:

� Virtual patients and simulation tools for training healthcare professionals

Administrative Efficiency:

Automation of routine tasks to reduce clinician burden

Patient Engagement:

�Tools that help patients understand and participate in their care

By 2025, over 75% of healthcare organizations have implemented or are planning to implement AI solutions to address clinical and operational challenges.

Source: Healthcare Information and Management Systems Society (HIMSS) 2025 Survey

AI in Medical Education: Overview

AI is revolutionizing how future healthcare professionals learn:

Virtual Patient Simulations

AI-powered virtual patients that respond realistically to student interactions

Adaptive Learning Systems

Personalized education pathways based on individual student performance

Performance Analytics

Real-time feedback on clinical reasoning and decision-making

"AI in medical education is not about replacing human instructors but augmenting their capabilities to provide more personalized, accessible, and effective learning experiences."

— Journal of Medical Education and Technology, 2025

AI for Learning History Taking

Virtual patients powered by AI provide medical students with realistic practice for history taking skills:

Realistic Conversations

Natural language processing enables human-like dialogue with virtual patients

Diverse Case Library

Exposure to wide range of patient presentations and medical histories

Immediate Feedback

AI analyzes questioning patterns and provides guidance on missed areas

Safe Learning Environment

Students can practice repeatedly without patient fatigue or time constraints

Research shows medical students using AI virtual patients for history taking practice demonstrate 32% improvement in comprehensive information gathering compared to traditional methods.

Source: Journal of Medical Education Technology, 2024

DDx demo

Improving Empathy & Bedside Manner

AI systems are helping medical students develop crucial interpersonal skills:

Real-time Feedback

AI analyzes tone, word choice, and body language during simulated patient interactions

Emotional Intelligence Training

Virtual patients display realistic emotional responses to different communication approaches

Personalized Coaching

AI identifies individual communication patterns and suggests specific improvements

"Students who trained with AI-enhanced empathy simulations showed a 42% improvement in patient satisfaction scores during clinical rotations."

— University of Arizona Health Sciences, 2024

AI systems like these provide safe environments for students to practice difficult conversations and receive objective feedback without risking patient discomfort.

Enhancing Diagnostic Skills

AI tools are transforming how medical students develop diagnostic reasoning:

Case-Based Learning

AI systems present diverse clinical scenarios with varying complexity and presentations

Interactive Feedback

Real-time guidance on diagnostic reasoning process and decision points

Pattern Recognition

Training on thousands of cases to improve identification of subtle clinical patterns

Studies show students using AI-enhanced diagnostic training demonstrate 40% improvement in accuracy compared to traditional methods.

Source: Journal of Medical Education Technology, "AI-Enhanced Diagnostic Training Outcomes," 2024

AI in Radiology: Overview

Radiology faces significant challenges that AI is helping to address:

122,000

Projected physician shortage by 2032, with radiology among the most affected specialties

30%

Annual increase in imaging studies in some healthcare systems

84%

Of radiologists report symptoms of burnout, partly due to increasing workloads

AI applications in radiology are expanding rapidly, from diagnostic assistance to workflow optimization and patient communication.

Sources: AAMC Report 2022, Radiology Business Journal 2024, American College of Radiology 2025

AI as a Copilot for Radiologists

AI systems are increasingly serving as intelligent assistants to radiologists:

Enhanced Detection

AI algorithms can flag potential abnormalities for radiologist review, reducing the risk of missed findings

Prioritization

Intelligent worklist sorting ensures critical cases are reviewed first

Automated Reporting

AI-assisted report generation with structured templates and measurements

Efficiency Gains

Studies show 22-29% reduction in reading time when using AI assistance

"AI doesn't replace radiologists; it augments their capabilities, allowing them to focus on complex cases and meaningful patient interactions."

— American College of Radiology, 2025

Worklist Prioritization

AI for Patient-Friendly Radiology Reports

AI is helping bridge the gap between complex medical terminology and patient understanding:

Original Radiology Text:

"Bilateral scattered fibroglandular densities with no suspicious masses, architectural distortion, or suspicious calcifications."

AI-Simplified Version:

"Your breast tissue appears normal. We did not find any concerning lumps or calcium deposits that might suggest cancer."

Google's RadExplain and similar tools use natural language processing to:

Translate medical jargon into plain language

Provide context for medical findings

Explain next steps in patient-friendly terms

Studies show patients using AI-translated reports have 43% better understanding of their condition and 37% reduced anxiety.

Source: European Society of Radiology, "ChatGPT-4o powers AI-generated simplified breast imaging reports," 2025

Rad Demo 1

Rad Demo 2

AI for Surgical Video Analysis

AI systems are revolutionizing how surgical videos are analyzed and utilized:

Automated Phase Recognition

AI identifies critical stages in surgical procedures, enabling precise analysis and comparison

Quality Assessment

Systems like MedGemini evaluate technical aspects such as Critical View of Safety (CVS) in laparoscopic procedures

Educational Tool

AI-annotated videos serve as powerful teaching resources for surgical trainees

Continuous Improvement

Analysis of surgical techniques across multiple procedures enables evidence-based refinement

Recent studies show that AI-powered surgical video analysis can identify critical safety steps with up to 84% accuracy, potentially reducing complications and improving outcomes.

Source: Journal of Surgical Innovation, 2025

Laparoscopic Cholecystectomy CVS demo

Improving Surgical Quality with AI

AI systems like MedGemini analyze surgical videos to assess quality metrics:

Critical View of Safety (CVS) Assessment

AI evaluates whether key safety criteria are met during laparoscopic cholecystectomy

Key Performance Metrics:

84%

Accuracy in detecting Critical View of Safety after training on 2,000 surgical videos

93%

Surgeon agreement with AI-generated surgical phase identification

40%

Reduction in adverse events when AI quality assessment is implemented

Source: Theator Inc. "How AI Can Improve Surgical Care" (2023), Journal of Surgical Innovation (2024)

AI for Surgical Training

AI-powered video analysis is revolutionizing surgical education:

Objective Assessment

AI provides standardized evaluation of surgical technique and performance metrics

Learning Curve Acceleration

Studies show 40% faster skill acquisition with AI-guided feedback compared to traditional methods

Deliberate Practice

AI identifies specific areas for improvement, enabling targeted skill development

Expert Emulation

AI compares trainee movements and decisions with those of experienced surgeons

"AI-guided surgical training has the potential to standardize surgical education and ensure consistent skill development across training programs."

— Journal of Surgical Education, 2024

Intuitive Surgical clip

SimNow2 clip

The Rise of Ambient Scribes

Ambient AI scribes are transforming clinical documentation:

Passive Listening

AI systems that automatically capture and transcribe doctor-patient conversations

Structured Documentation

Converts natural conversation into organized clinical notes integrated with EHR systems

Time Savings

Physicians using ambient scribes report

2-3 hours

saved daily on documentation

By 2025, over 40% of healthcare organizations have implemented or are piloting ambient scribe technology to reduce clinician burnout and improve patient interaction.

Source: JAMA Network Open, "Physician Perspectives on Ambient AI Scribes," 2025

NEJM Catalyst, "Ambient Artificial Intelligence Scribes: Learnings After 1 Year," 2025

Leading Ambient Scribe Solutions

Abridge

Real-time transcription with "Linked Evidence" technology

Valued at $5.3 billion (June 2025), used in 7,000+ healthcare facilities

Nuance DAX (Microsoft)

Dragon Ambient eXperience with deep EHR integration

Deployed in 150+ health systems with Epic integration

Ambience Healthcare

AI platform for documentation, coding, and clinical workflow

Selected by Cleveland Clinic after extensive evaluation process

Doximity & OpenEvidence

Free AI scribe (Doximity) and medical information platform

OpenEvidence used by 40% of US physicians, valued at $3.5B

"Ambient AI scribes have shown positive impacts on physician workload, work-life integration, and patient interactions in multiple studies."

Sources: JAMA Network Open (2025), Fierce Healthcare (2025), LinkedIn posts by Rik Renard (2025)

AI for Administrative Tasks

AI is streamlining administrative workflows in healthcare:

Scheduling & Registration

Intelligent systems that optimize appointment scheduling and automate patient registration

Billing & Coding

AI-powered coding assistance that improves accuracy and reduces claim denials

Patient Communication

Automated appointment reminders, follow-ups, and medication adherence checks

Data Exchange

Streamlined requests for patient records between healthcare providers

Healthcare organizations implementing AI for administrative tasks report an average

30% reduction

in administrative burden and

25% increase

in staff productivity.

Source: Healthcare Information and Management Systems Society (HIMSS), 2025

AI in Revenue Cycle Management

AI is transforming healthcare revenue cycle management:

Claims Processing

AI predicts and prevents claim denials, reducing rejection rates by up to 30%

Clinical Documentation

AI-powered coding assistance ensures accurate and compliant medical coding

Prior Authorization

Automated workflows reduce authorization processing time from days to hours

Predictive Analytics

AI forecasts payment timelines and identifies optimization opportunities

Healthcare organizations implementing AI-powered RCM solutions report an average 27% reduction in days in accounts receivable and 35% improvement in clean claim rates.

Source: Healthcare Financial Management Association (HFMA), "AI in Revenue Cycle Management," 2025

Medicine is not black and white

AI alone vs (AI + Doctors) vs Doctors alone

AI for Patient Engagement

AI is empowering patients to become more active participants in their healthcare:

Health Literacy

AI tools translate complex medical information into accessible, personalized explanations

Treatment Adherence

Smart reminders and personalized education improve medication compliance by up to 40%

Pre-visit Preparation

AI assistants help patients formulate questions and organize concerns before appointments

Ongoing Monitoring

AI-powered apps track symptoms and provide actionable insights between clinical visits

"When patients arrive with AI-assisted research, it creates opportunities for more meaningful and productive clinical conversations."

— JAMA, "When Patients Arrive With Answers," 2025

Hippocratic AI voice agent clip

​

Leveling Information Asymmetry

AI is transforming the traditional knowledge gap between patients and providers:

"When patients feel unheard, arming themselves with knowledge becomes a strategy to be taken seriously."

— JAMA, "When Patients Arrive With Answers," 2025

AI tools provide patients with medical information previously accessible only to clinicians

Shifting from "doctor knows best" to shared decision-making model

Patients arrive better prepared for meaningful dialogue about their care

The Atlantic notes: "At its best, medicine will be a process of shared decision making, and doctors need to be prepared." This requires both common knowledge and clear communication frameworks between patients and providers.

Sources: JAMA, "When Patients Arrive With Answers," 2025; The Atlantic, "When Patients Do Their Own Research," 2024

AI can (rarely) make mistakes

Is AI making doctors better ?��(can it, unintentionally, make them worse ?)

Future Directions and Challenges

As AI continues to transform healthcare, several key challenges must be addressed:

Privacy and Security

Ensuring patient data protection while enabling AI systems to learn from clinical data

Ethical Considerations

Addressing bias, transparency, and accountability in AI healthcare applications

Clinical Validation

Establishing rigorous standards for evaluating AI performance in real-world settings

Workforce Adaptation

Training healthcare professionals to effectively collaborate with AI systems

The future of AI in healthcare will depend on thoughtful integration that enhances human capabilities while addressing these challenges. Regulatory frameworks are evolving to keep pace with technological innovation.

Source: World Health Organization, "Ethics and Governance of Artificial Intelligence for Health," 2025

Conclusion

AI agents are transforming healthcare across multiple domains:

Medical Education:

Enhancing training through virtual

patients, empathy development, and diagnostic skill

building

Clinical Practice:

Supporting radiologists, improving

surgical quality, and reducing documentation burden

Healthcare Operations:

Streamlining administrative

tasks and optimizing revenue cycle management

Patient Experience:

Empowering patients with

information and improving engagement in care

decisions

"AI in healthcare is not about replacing human expertise, but augmenting it—creating a future where technology enhances the human elements of care."

Thank you for your attention

August 2025

Technology & ai is not a panacea !

​

Where to learn more ?

​

​

​

http://mlformds.com/

​

​

Thank you for your attention !�Questions ?

@inventcures �@KCDH_A��on Twitter/X