316: Punit Soni, CEO of Suki, On Healthcare AI Revolution, Voice Assistants, and Empowering Clinicians
AI and the Future of Work
Deep Dive
What is Suki, and how does it aim to revolutionize healthcare?
Suki is a voice-based digital assistant designed to help clinicians by automating administrative tasks like clinical documentation, coding, and order entries. It aims to make healthcare technology assistive and invisible, allowing clinicians to focus more on patient care rather than administrative burdens. By leveraging AI and voice interactions, Suki seeks to reduce the time doctors spend on non-clinical tasks, which currently takes up 30-40% of their time.
Why did Punit Soni choose healthcare as the focus for Suki?
Punit Soni chose healthcare because it is a domain with sophisticated users, a significant administrative burden, and repetitive tasks. He saw an opportunity to apply AI and voice-based interactions to solve these challenges. Healthcare also lacks a dominant tech company due to its fragmented nature, making it ripe for innovation. Suki's goal is to democratize healthcare tech by creating a unified experience that addresses multiple pain points for clinicians.
How does Suki ensure trust and transparency in its AI-driven healthcare solutions?
Suki ensures trust and transparency by clearly defining its role as an assistant, not a replacement for clinicians. It provides tools like the 'transcript view,' which allows doctors to trace how clinical notes were generated from patient conversations. Additionally, Suki emphasizes responsible AI practices, including double sequential de-identification of personal data to protect patient privacy. All outputs are under clinician oversight, ensuring that doctors approve and validate the AI's work.
What are the primary use cases for Suki in clinical settings?
Suki's primary use cases include clinical documentation, coding, order entries, and providing contextual information like patient summaries and medication details. It also assists with data retrieval, such as pulling up vaccination records, and can generate insights like plotting A1C levels over time. Over time, Suki aims to become a comprehensive assistant that handles scheduling, patient summaries, and other administrative tasks, allowing clinicians to focus on patient care.
How does Suki handle the challenge of protecting patient data while training its AI models?
Suki uses double sequential de-identification to strip out personally identifiable information (PII) and protected health information (PHI) before any data is used to train its AI models. This ensures that no sensitive patient data is exposed. The company also employs a thoughtful architecture design to overlay contextual patient information securely without compromising privacy. This approach allows Suki to leverage data insights while maintaining strict confidentiality.
What is Punit Soni's vision for the future of healthcare with AI?
Punit Soni envisions a future where AI makes healthcare more decentralized, efficient, and accessible. He believes AI will act as a scalable assistant to clinicians, enabling them to care for more patients and reducing the global shortage of healthcare professionals. In this future, patients will own their health data, and AI agents will facilitate seamless interactions between patients and doctors. Soni predicts that AI will lead to longer, healthier lives and a healthcare system that is less frustrating and more empowering for all stakeholders.
What skills does Punit Soni believe are future-proof in the age of AI?
Punit Soni emphasizes the importance of learning math, philosophy, and history as foundational skills for the future. He also encourages creativity and adaptability, as new technologies like AI and robotics will create entirely new industries and opportunities. Soni believes that understanding the past and thinking critically will be crucial in navigating the rapid changes brought by AI, enabling individuals to contribute meaningfully in fields like advanced manufacturing, space exploration, and healthcare.
Shownotes Transcript