This is part of a series of conversations with CBS alumni building or leading businesses tackling the most pressing AI-related challenges and opportunities. Michael Kopko is the CEO of Pearl Heath, an AI-powered platform for care teams’ workflows.
What is the challenge Pearl Health is addressing and how has that vision evolved as AI has entered the picture?
Pearl Health was founded to empower independent physicians through data and value-based care. The founding vision of Pearl Health—empowering independent physicians through data and value-based care—has only accelerated with the advent of AI. Before the release of ChatGPT, many investors thought our vision of changing healthcare behavior without a heavy labor model was unachievable. Now, this is the expectation.
Given all of the data that we now have, almost everything in senior care is predictable and therefore preventable. Better still, we can automate the actions required to prevent those potential issues occurring.
As CEO, how do you think about setting ethical and operational guardrails for AI—balancing speed of innovation with patient safety and regulatory complexity?
We have to balance speed of innovation with patient safety, and safety is critical when dealing with people's healthcare. However, humans make errors, too. That leads us to this question of, what are you missing when you make mistakes and what's the level of the miss? Let’s take our approach to predicting hospitalization events. If we're off on that prediction, there are two types of misses we could have:
We could predict someone going to hospital who doesn't in fact go. What's the harm in that? We call someone and ask for an appointment with their primary care. Very low downside to that.
The other potential error is that we don't predict someone who's about to go to hospital. This would be disappointing but the patient is no worse off than pre-Pearl.
On the regulatory side, I do have concerns. State-based regimes and rules designed to “protect” consumers now significantly increase the cost of care and slow down market entry for critical care solutions. Although I understand the need for these regulations, I think we need to make some improvements here and dramatically simplify these regimes. The lawyers and administrators capture far too much of our healthcare dollar these days.
How do you see the relationship between physicians and AI evolving—not just in workflows, but in mindset and trust—over the next five years?
I believe that physicians won't be replaced by AI, but physicians will be replaced by physicians who see AI as an enabler and thought partner. I foresee a future where providers supervise, train, and sharpen AI algorithms, effectively giving them the power of many physicians. This shift will change healthcare delivery from an increasingly outdated one-to-one visit modality to a one-to-many model, allowing physicians to engage with patients through simultaneous, agentic responses and system management, not requiring them to be physically present in all circumstances.
Historically one of the main hurdles with more widespread adoption of value-based care (VBC) has been identifying and tracking the right measures of quality. What has the impact of AI been on addressing these challenges?
I think the current VBC quality measures are often too process-driven and are orthogonal to true quality, where targets are set just to be met. Real VBC quality should focus on cost and outcomes, and services should become commoditized to reduce the wild variance in pricing for similar care, like surgery. AI will make achieving lower costs and higher quality more quickly attainable by cutting labor costs and intermediaries and enabling predictive capabilities to proactively deliver healthcare without patients needing to ask.
Looking back to your early career, what experiences most shaped how you think about technology, data, and human decision-making today?
My early entrepreneurial experiences significantly shaped my views on technology and data. I learned the importance of timing in sales from running my college dorm business, Dormaid, where all sales effectively happened within two weeks.
I also learned from an early age to integrate technology into what I was offering. As a third grader, selling magazines, I used spreadsheets to do pricing and get price lists out there. And with Dormaid, it was about using the internet to sell services. Now that’s commonplace but, back then, it was very rare, particularly for selling to college students.
You have to make sure you’re catching what is coming in the next five to ten years and really be intentional about where you think the world is going versus being backward looking. You then integrate those arcs into your strategy and offering and position yourself as best you can for distribution and continuous improvement.
What advice would you give to fellow alumni or aspiring founders who want to tackle big, complex problems in healthcare or technology?
I have four key lessons for future leaders:
First, you need to stay flexible. I'm worried we're becoming very herdlike as leaders and it's important to stay independent, objective, and have all of your mental models be updatable by the data you're seeing, not based on your prior strongly held opinions. Second, you need to have a high tolerance for pain, or level of grit, because accomplishing big things requires enduring losses and setbacks. Third, your peer group is really critical. It not only defines you, but it defines the standards that you set and the ambition that you target. Surround yourself with people that you admire and absorb their standards and ambition.
Finally, enjoy the ride.
