Work that went deep.

The Problem

Real AI ambition existed across the organization, but without a shared way to evaluate where it's most likely to work. The result: initiatives ran in parallel with no common language for comparing them. No framework for weighing time, people, or capital against any given project was actually expected to return.

The Approach

The first step was establishing that common language: a lightweight but quantitative prioritization scheme that could hold both internal initiatives (efficiency gains, workflow augmentation) and external ones (product design, new AI-native development) in the same conversation. From there, a leadership structure and operating model were stood up to execute against it.

The Outcome

• ✓Cross-enterprise AI opportunity assessment performed
• ✓Prioritization frameworks, operating model, and leadership structure in place
• ✓Prioritized AI roadmap designed for 1- and 3-year financial targets delivered

The Problem

Real-time clinical decisions about disease progression, mortality risk, and care transitions were being made without patient-specific intelligence — just clinical intuition. The gap was in getting that intelligence in a usable form, at the point of care.

The Approach

A suite of patient-specific risk models was built and validated across several domains: secondary hypertension, behavioral health acuity, and cardiac health. Each was embedded directly into the EHR, not as a separate tool clinicians had to seek out. Transparency and safety monitoring were built in from the start, not added later.

The Outcome

• ✓7M+ patients reached by AI products delivered
• ✓Improved clinical adoption and patient outcomes
• ✓Published, peer-reviewed models
• ✓Nationally recognized, award winning AI products

The Problem

Denied insurance claims were eating margins and demanding time from skilled labor. Staff were writing appeal letters by hand, slow and inconsistent, consuming time that could have gone elsewhere. The revenue loss was substantial. So was the cost of the people doing the work.

The Approach

Product managers, solution engineers, and data scientists were embedded directly with the revenue cycle team and designed a human-in-the-loop AI solution for generating appeals letters. The result was more timely appeals, a higher appeal rate, reduced write-offs, and optimized staff time. The AI didn't replace the people doing the work, but rather amplified their impact.

The Outcome

• ✓Tens of millions in recovered revenue
• ✓Thousands of labor hours saved annually
• ✓Redesigned workflow with sustained team adoption