Why an AI-Powered Semantic Layer is the Key to Unlocking Healthcare's Data-Driven Future

Early Career: Analytics as Justification

I started my career at the intersection of capital decisions and patient care, helping hospitals justify and operationalize major purchases like the da Vinci and Mako surgical robots. 

At this stage, my job was all about justification, with questions like “Is there enough surgical volume to support a $1.5 million robot?” We had to balance a massive business decision with genuine patient needs, and we used procedural data to prove the ROI. (It’s amazing to see how far things have come; I recently visited Johns Hopkins and was able to test-drive a da Vinci myself, as they've become the standard of care.) 

A Shift to Operations: Tracking the Patient Journey

Later, I moved into referral management, working with a tool called Crimson Medical Referrals. This is where I first saw the power of operational analytics. With Strategy (then MicroStrategy), we tracked referral leakage and continuity of care, asking questions like “Were primary care physicians referring patients to in-network specialists?” 

This operational data ensured patients received timely, coordinated care, an approach that was both clinically beneficial and financially sustainable for the health system.  

But even then, we were mostly looking in the rearview mirror. 

Today: Building the Proactive, AI-Ready Foundation

Now, working at Strategy, I've seen the full picture. Data isn’t just about reporting past events or justifying capital; it’s about proactively shaping the future of healthcare delivery. 

Health systems are now leveraging their data to drive measurable outcomes; a national provider we work with identified $10 million in cost savings by optimizing surgical procedures alone. 

But to do this, you must solve the fundamental problem that has plagued healthcare for decades: fragmented data. This is where Strategy Mosaic, our AI-Powered Universal Semantic Layer, becomes the critical foundation. 

The Mission-Critical Need for a "Single Source of Truth"

The concept of a Single Source of Truth (SSOT) is the idea that every decision-maker, from a clinician to an accountant, accesses the same governed metric, regardless of the tool they use. 

To understand this, you have to understand the semantic layer. 

At its simplest, a semantic layer like Strategy Mosaic turns data complexity into business clarity. It sits on top of all your disparate systems (EHRs, billing, HR, etc.) and translates technical complexity into consistent, user-friendly business concepts. So instead of seeing LU_LOS in a report, you see Patient Length of Stay, pre-calculated for you. 

Think of it as the master recipe book and well-stocked kitchen. All departments start with the same pre-approved ingredients and follow the exact same recipe. For example, when a clinician pulls up “30-day readmission rate,” they're not seeing different calculations based on their department. The CFO looking at cost metrics and the Chief Medical Officer examining quality indicators are working from the same foundation—eliminating the “my numbers vs. your numbers” debates that plague healthcare decision-making. 

In healthcare, this isn't a nice-to-have; it's mission-critical: 

1. Regulatory, Governance & Financial: Health systems must report metrics to numerous bodies, from CMS to Leapfrog. A SSOT guarantees that "readmission rate" is defined consistently everywhere, avoiding fines. Just as critically, it provides a centralized and auditable layer of data governance. This gives you full visibility into who is accessing what data, which is essential for protecting PHI and ensuring HIPAA compliance. 

2. Clinical Trust: It eliminates duplicate or conflicting patient records, reducing the risk of errors in diagnostics, medication dosage, and care planning. 

3. Operational Efficiency: Standardizing data like provider credentialing streamlines administrative processes, reducing bottlenecks so patients can get appointments faster.

How Mosaic Delivers a True, Unified Layer

What sets Mosaic apart is how it delivers this single source of truth in a way that solves real-world problems. 

• It’s Portable and Vendor-Agnostic: This addresses a pain point I experienced working in referral management. I would build a beautiful Strategy dashboard, but analysts at various hospital groups remained Excel die-hards. The moment they downloaded the data, it became disconnected, outdated, and produced conflicting numbers. 

Mosaic solves this. The source of truth lives in the semantic layer, and analysts can connect Excel or Power BI directly to that governed model. When a metric is updated, it updates instantly for everyone, in every tool. 

• It Makes Building Easy (No YAML Required): Though Strategy has 25+ years of experience in semantic layers, I'll be honest; it used to be hard to build.  

But I would have loved Mosaic as an analyst. Our new AI-powered Mosaic Studio lets business users build their own data products with natural language. It helps you create data relationships, joins, metrics, and attributes. This opens the door for analysts to remix and publish new data products themselves. 

• It's Future-Proof: Since Mosaic’s intelligence layer is decoupled from the underlying data warehouse, you control the technology. If you migrate from Snowflake to Databricks, you simply repoint Mosaic. All your dashboards, reports, and AI applications continue working. 

Proof in Practice: Better, More Equitable Maternal Care

The best part is seeing how this foundation enables better patient outcomes.  
A powerful example comes from our customer, OBHG Management Services LLC. As a parent myself, I love this story. 

OBHG provides OB/GYN hospitalist programs to over 175 partner hospitals. They use a Centralized Operations Dashboard, powered by Strategy, to review key metrics across all their programs. 

In October 2024, they released a report that analyzed maternal health outcomes across racial groups. They found that their hospitalists are able to deliver unbiased care, with outcomes consistently outperforming national averages for all racial and ethnic identities. 

This is only possible because Strategy provides the row-level security and SSOT needed to centralize extremely sensitive, real-time patient data. 

And it’s not just clinical. By using Strategy's governed calculations, OBHG streamlined its revenue cycle, reducing its monthly billing cycle from 20 to 17 days, and slashing manual reconciliation efforts by 60%. 

The Future: An AI-Ready Foundation for Total Health

I'm most excited about the possibilities of Generative AI and integrating data from all sources to get a better picture of total health. When I had gestational hypertension, I had to retell my story to every single provider—from the phlebotomists to my son’s pediatrician. My stress would have been so much lower if my watch data, last visit details, and labs were all unified. 

We're now moving beyond static dashboards to AI Agents that can answer complex business questions. But the biggest risk in using AI in healthcare is an AI hallucination—an answer based on unreliable or biased data. You cannot have an AI hallucinate a patient's treatment plan. 

This is why Mosaic is so critical. It is intentionally designed to be the foundational AI Data Layer. It acts as a crucial barrier, providing large language models (LLMs) with context-rich, governed metrics—without ever exposing raw, patient-identifiable data (PHI). 

It solves interoperability and data-quality challenges first, guaranteeing that the AI's answers are accurate, ethical, and explainable. By transforming fragmented data into reliable enterprise intelligence, we can finally empower providers to stop debating which number is correct and start acting on insights to improve patient outcomes. 

As healthcare continues its data transformation journey, the organizations that thrive will be those that build on a foundation of trusted, unified data. The clinicians, administrators, and patients I worked with years ago deserved nothing less. Today, we finally have the technology to deliver on that promise.