Earlier this month, at a packed event in New York City, KAID Health CEO Kevin Agatstein shared his insights on the transformative power of AI in healthcare. Speaking to an engaged audience, he outlined key lessons from his journey in building and commercializing AI-driven solutions—insights that apply not just to KAID Health but to any AI technology aiming to succeed in the healthcare industry.
Here are the five essential takeaways from his talk.
1. AI Must Address Both Clinical and Financial Needs
Healthcare AI cannot succeed unless it delivers tangible value to both patients and providers. KAID Health started with a simple yet powerful mission: to use AI to summarize patient medical records, reducing the burden on physicians and freeing them to focus on patient care.
However, success in healthcare AI requires more than clinical benefits—it must also drive financial value. Our earliest customers, like The Villages Health, made it clear: buried in electronic medical records were millions of dollars in unrealized revenue due to incomplete coding. By adapting our AI to address risk adjustment and coding inefficiencies, we created a solution that not only improves patient care but also helps providers get paid accurately.
The lesson? AI must solve a real problem—one that impacts both clinical workflows and financial sustainability.
2. It Takes GREAT AI to Add Value—Invest in Model Performance and Validation
AI in healthcare is only as good as its data, training, and validation. Early on, KAID Health experimented with third-party AI models. One initial test flagged “Precocious Puberty” as a common diagnosis among elderly patients—clearly a misfire.
To build an AI solution that physicians can trust, we invested heavily in model performance and validation. We insourced AI development, refined our models, and partnered with UCSD Health’s AI lab to ensure rigorous academic validation. The result? Peer-reviewed studies demonstrating that our AI can outperform physicians in summarizing key aspects of patient charts.
The takeaway? Poor AI is worse than no AI. In healthcare, precision matters—invest in the right data, model training, and validation.
3. AI Alone Does Not Make a Product
Even the best AI is just one piece of the puzzle. While our AI could dramatically improve medical coding efficiency, we quickly learned that most healthcare organizations need more than just an algorithm—they need a complete, end-to-end solution.
To meet this need, we built an entire AI-powered medical coding workflow on top of our existing platform. We also hired and trained medical coders to provide AI-assisted coding services for organizations that lacked the internal capability. Some of our customers now use KAID Health purely as a technology platform, while others rely on our full-service coding support.
The lesson? AI is not the product—the complete solution is. Build with your customers, not just for them.
4. It Takes Partners to Get to Market
Healthcare is an industry with long sales cycles and complex decision-making processes. Rather than selling directly to individual providers one by one, we accelerated our growth by partnering with organizations that already have an installed base of customers.
Today, KAID Health works with leading ACOs, payer organizations, and healthcare service providers to integrate our technology into their existing solutions. These partnerships have allowed us to scale faster and deliver value at a much larger scale than we could have alone.
The takeaway? In healthcare AI, go-to-market strategy is just as important as the technology itself. Find the right partners to expand your reach.
5. In Applying AI, Costs Matter
While AI is often seen as a cost-saving technology, running AI models—especially high-performance ones—can be expensive. To deliver real-world impact, we must balance AI sophistication with cost efficiency.
For example, KAID Health doesn’t rely on a single “rock star” AI model for every task. Instead, we use lighter, more cost-effective models for simpler processes while deploying high-powered AI models only where needed. This approach allows us to summarize medical records efficiently—at a cost of under 5 cents per page—without sacrificing quality.
The lesson? AI must be cost-effective at scale. Smart AI implementation requires a business mindset, not just technical brilliance.
Final Thoughts
AI has the power to transform healthcare—but only if it is practical, validated, and implemented in a way that makes both clinical and financial sense. From solving the right problems to ensuring cost efficiency, these lessons are essential for any AI company looking to succeed in this space.
At KAID Health, we’ve applied these insights to build an AI-driven medical coding platform that improves both clinical workflows and financial performance. As healthcare AI continues to evolve, these principles will remain foundational for driving meaningful change.
Would you like to discuss how AI-powered medical coding can enhance your organization’s efficiency? Let’s talk.