A Generative AI Primer for Healthcare Business Leaders

According to Gartner, 80% of health system CIOs are actively exploring Generative AI use cases, so on this episode of The Innovation Engine, we break down the basics that everyone in health IT should know about the technology. 

We cover three main topic areas that are vital for healthcare leaders to understand when exploring Generative AI solutions: accuracy, cost, and privacy.

Pankaj Chawla, 3Pillar’s Chief Innovation Officer, and David Evans, 3Pillar’s Director of Global Innovation, join the show to talk with 3Pillar’s Healthcare Industry Lead Steve Rowe. They give listeners and healthcare leaders a look at how they can and should harness GenAI to elevate the patient experience.

On the accuracy front, we cover topics including why retrieval-augmented generation (RAG) is the most common architecture to be placed over top of LLMs, as well as the importance of data engineering and being able to provide accurate data sources so an LLM can use natural language processing to deliver the right outputs.

When it comes to managing costs, David and Pankaj explore a number of ways to ensure you manage AI costs effectively. While companies can license seats for a broad range of business users to any of the most popular LLMs, there are significant cost advantages to creating homegrown alternatives. An internal chat application that David and the Innovation Lab team have developed at 3Pillar is a good place to look for an example.

Rather than paying $20 per user/month for dozens of users, the team has provided a viable solution for roughly 100 users at a cost that is orders of magnitude lower than it would have been using paid versions of ChatGPT, Microsoft Copilot, and other similar tools.

What about the data privacy side of the equation, which is always a concern for healthcare leaders due to regulations like HIPAA and the sensitivity of the data they possess? Pankaj and David cite some of the same practices that companies have adopted for years to secure their cloud applications (private, secure infrastructure) as a successful approach to ensuring data privacy when working with AI. This can include provisioning your own GPU for even greater security, not to mention additional cost control.

Listen to this episode of The Innovation Engine on Apple Podcasts, Spotify, or via the embed below.

You can watch the full episode via the YouTube embed below.

• 00:00 – Intro to GenAI in Healthcare
• 1:53 – Why hallucinations happen
• 3:56 – Preventing hallucinations
• 11:01 – Cost concerns
• 13:37 – Keeping GenAI cost effective
• 19:25 – Data + privacy concerns
• 21:07 – Protecting data
• 25:30 – Data architecture for GenAI

• Read 3Pillar’s white paper: “How Can You Keep Your AI Costs From Spiraling Out of Control?”
• Read about the internal chatbot that JPMorgan Chase developed for employees
• Learn more about 3Pillar’s GenAI Strategy 
• Email Steve Rowe or connect on LinkedIn
• Email Pankaj Chawla or connect on LinkedIn
• Email David Evans or connect on LinkedIn
• Read Steve’s latest blog posts for 3Pillar