For much of my career, I was immersed in the world of healthcare technology, working with some of the largest organizations like Kaiser Permanente, UnitedHealthcare, and Optum. My job was to build and transform healthcare platforms—no small task in an industry as complex and data-heavy as healthcare. The more I worked on these projects, the more I encountered the same frustrating reality: the healthcare system, particularly in the U.S., is deeply rooted in document-based, human-centered processes, and the technology available simply wasn’t built to solve its unique challenges. Despite the rapid advances in machine learning, natural language processing (NLP), and business intelligence (BI) over the last few decades, most of the tools and platforms available were, at their core, generic. I would be in discussions with vendors from companies like Snowflake, Databricks, AWS, GCP, and Azure, who were eager to sell their analytics capabilities. Yet, as a buyer tasked with solving specific, vertical industry problems in healthcare, I was continually met with disappointment. These platforms were designed for broad, horizontal applications. They offered basic, generalized tooling and infrastructure but left the burden of building healthcare centric solutions entirely on us. And that’s the crux of the issue. These technology companies expected us to use horizontal platforms to solve the highly specialized and complex problems that healthcare presents. They made little effort to understand the intricacies of healthcare workflows, the unique nature of our data, or the stringent compliance and governance requirements we face. Their tools were generic, and we were left to fill in the gaps ourselves.
The Unmet Needs of Healthcare
Healthcare payers and providers are not just another vertical. They operate in a complex, regulated environment, filled with sensitive, unstructured data. The workflows—from claims processing to clinical decision support—are intricate and deeply intertwined with compliance, privacy, and security protocols. Any solution must be able to handle these requirements natively, not as an afterthought. The platforms I was being offered failed to address this. Healthcare wasn’t even an afterthought. It was left entirely to us, the buyers, to figure out how to mold these horizontal tools into something usable. We had to bolt on healthcare-specific workflows, train our teams to manage the tools, and often spend hundreds of millions of dollars and years of effort just to make these platforms functional in a healthcare context. This resulted in systems that were outdated before they even went live, incapable of solving the true challenges we were facing. Consider this: one project aimed at integrating member data, processes, and analytics was estimated to take seven years and cost $1.2 billion. Another involved a provider analytics system that had been in use for 15 years, which was nearly impossible to replace without astronomical costs. Even our clinical decision support systems were sluggish, relying on batch processes that took over 24 hours to complete. It felt like a never-ending cycle of slow, incremental progress.
The Paradigm Shift: Generative AI and LLMs
Everything changed when I started working with large language models (LLMs) and Generative AI. These technologies represent a true paradigm shift—a step change in what’s possible for healthcare. After decades of feeling stuck with outdated, cumbersome systems, here was a technology that could solve the problems that had plagued us for so long. For the first time, I saw the potential to tackle tasks that only humans could perform before. LLMs can handle unstructured data, which makes up the vast majority of healthcare’s valuable information. From clinical notes to patient histories and insurance forms, LLMs can process these documents quickly and accurately, unlocking insights that were previously hidden in text or PDFs. This was something that no other technology before had been able to achieve effectively. Suddenly, the idea of automating complex, multi-step processes became real. Generative AI models could not only read and interpret unstructured documents but also generate human-like responses and insights. Whether it was summarizing patient records, predicting disease outcomes, or streamlining prior authorizations, Generative AI was opening doors that rules-based systems and BI tools never could.
Why Healthcare Needs Its Own AI Platform
As powerful as LLMs are, the real challenge lies in integrating them into the healthcare ecosystem. Healthcare organizations don’t just need general-purpose AI models—they need AI models that are pre-trained on healthcare data, understand the complexities of medical terminology, and are built with governance, compliance, and security at their core. This is why healthcare deserves its own AI platform. Healthcare-specific AI platforms must be designed from the ground up to meet the unique needs of the industry. They should come equipped with pre-trained models that understand common healthcare tasks like medical coding, claims adjudication, and clinical decision support. These platforms need to seamlessly connect to healthcare data systems, including electronic health records (EHRs), claims databases, and imaging systems, all while complying with the stringent regulations that govern patient data, such as HIPAA and GDPR.But it doesn’t stop there. Healthcare AI platforms must also integrate into the existing workflows of payers and providers. Healthcare organizations have already invested heavily in systems like Epic, Cerner, and other EHRs. Any new AI platform needs to fit into these environments, enhancing them rather than replacing them. A generic AI platform cannot meet these needs. Healthcare requires a solution that understands the intricacies of the industry, from the governance and compliance requirements to the unstructured nature of its data. Without this, organizations will continue to struggle with fragmented, siloed systems that don’t deliver on their promise of efficiency and cost reduction.
My Frustrations as a Buyer
I remember the countless vendor meetings where I would sit and listen to pitches from companies that had built great tools for general-purpose analytics but had no idea how to apply them in healthcare. They’d show me data lakes, data warehouses, and analytics platforms that were impressive in their own right—but completely disconnected from the realities of healthcare. They didn’t get it. They didn’t understand the urgency of solving healthcare’s real problems—problems like reducing the administrative burden that costs the U.S. healthcare system $1 trillion annually. They didn’t grasp the need for platforms that could parse clinical notes in real-time, predict outcomes based on unstructured data, or streamline prior authorizations using natural language processing. Instead, they wanted to sell me horizontal platforms and left the hard work of customization to us. We had to build everything ourselves—spending years and millions of dollars just to get a system that was “good enough” to function. It was incredibly frustrating.
A Vision for the Future
It’s time for healthcare to have its own AI platform—one that understands the unique challenges of the industry and provides solutions tailored to its needs. This is the vision behind Penguin AI. We’re building a healthcare-specific AI platform that doesn’t just offer tools but delivers ready-to-use solutions designed to integrate with the systems payers and providers already rely on. With pre-trained healthcare language models, our platform can handle the complex tasks that healthcare requires—everything from prior authorization to claims processing to clinical decision support. We connect directly to healthcare data systems, ensure compliance with the strictest governance requirements, and offer seamless integration into existing workflows and applications. Healthcare deserves better than generic tools. It deserves a platform built specifically for its needs—one that can unlock the full potential of AI and transform the way we deliver care. At Penguin AI, we’re not just building technology; we’re rethinking how technology can serve healthcare. By focusing on the unique requirements of the industry, we aim to empower healthcare organizations to embrace AI with confidence, knowing that they have a partner who understands their challenges and is committed to solving them. The future of healthcare is not just about using AI—it’s about using the right AI. It’s about using a platform that can truly meet the needs of healthcare payers and providers, and that’s exactly what we’re delivering.
.svg)