60%
Traditional data engineering
Clean interfaces between Epic, Innovaccer, and Salesforce. The unglamorous plumbing that makes the rest of the stack honest.
A working note, close to home
A read on where AI fits inside Baptist Health’s stack, and where it does not.
Epic is mid-flight. Innovaccer carries population health. Salesforce Data Cloud is building the patient view. New AI features are arriving inside the EHR and the pharmacy every quarter. This page is one quiet read on where those layers meet, and where the governance seams open up.
The moment
Several layers, one governance question.
Baptist Health is standardizing onto Epic while Innovaccer carries population health, Salesforce Data Cloud builds the patient view, and new AI features arrive inside the EHR and the pharmacy every quarter. That is a lot of layers to govern at the same time, and governance is the part that gets quiet until it isn’t. The team has written thoughtfully about the promises and pitfalls of AI in medicine. The work below is how we would help make sure the promises ship and the pitfalls get caught early, close to home, close to the clinician.
The Eduba frame
Most of what a system this size needs from AI is not AI.
Our job is to look at a specific workflow at Baptist and decide which layer each step belongs on, before any new model is procured. We call it computational orchestration. It is quiet work and it saves real money.
30%
Rule-based logic
Codifying how Baptist already practices care. Predictable, auditable, defensible when a regulator or a clinician asks.
10%
Language models and agents
Ambient documentation, inbox triage, cross-institute referrals. The places where models actually earn their keep.
The relevant case study
KPMG UK, one of the Big Four.
We did a version of this with KPMG UK (Big Four): forty-plus executives in a regulated-industry setting, walked through a working frame for how to think about AI procurement and governance in a professional services environment. The engagement was measured, not splashy, and it gave the leadership team a shared vocabulary for evaluating vendor claims.
The same shape of work fits Baptist Health’s current moment: an enterprise in the middle of an Epic migration with several AI vendors arriving at once, where a shared governance frame across the CDIO office, the innovation unit, and the service-line medical directors would pay for itself quickly.
- 40+ executives trained, leadership-level audience
- Regulated industry, methodology-first posture
- Shared vocabulary for evaluating vendor claims
The relevant paper
Ethics Engine.
Our Ethics Engine paper is a psychometric assessment tool for evaluating ideological and moral patterns in LLMs. It is the kind of rigor a health system with a published AI-in-medicine posture will recognize.
The companion methodology, Interpretable Context Methodology, is a fit for a later conversation about clinical documentation architecture when the Epic build stabilizes.
Partnership line
Eduba partners with NLP Logix for work that sits below the orchestration layer. NLP Logix has been in machine learning since 2011 and runs over 150 data scientists.
Next step
Thirty minutes with Matt. Bring one workflow that keeps a clinician at the keyboard past their shift.
We will walk through how to decide which layer each step belongs on, live. Scoped, quiet, close to the clinician. No pitch deck.