Hey Reader,

The Missing Layer Between AI Hype and AI Success

How Precursor’s Digital Shadow Prevents the 95% Failure Rate

Let that sink in. Ninety-five percent. Billions invested. Almost nothing to show for it.

But here's what the research reveals: The AI itself usually works fine. The technology isn't the problem. The problem is that AI is deployed to processes that are more complex than they appear. They contain hidden dependencies, informal workarounds, and tribal knowledge outside formal documentation.

Robert Adamson, CIO of RWJBarnabas Health, captured the root cause perfectly:

Hospitals are treating AI as the strategy instead of understanding what they’re trying to accomplish first. And without that operational clarity, AI can’t help—it can only magnify confusion.

What AI Failure Actually Looks Like

Take asset management AI. A hospital deploys a $450K system to solve their “missing” IV pump problem. The AI analyzes RTLS data and predicts equipment needs based on census and procedure schedules. Six months later, nurses are still searching for pumps.

Why? The AI was trained on formal checkout procedures, but 60% of pumps move without being checked out because the formal process takes 8 minutes. The AI predicts demand beautifully based on data that doesn’t reflect reality. Nobody mapped the actual workflow. Nobody captured that night shift “borrows” pumps without documentation because the approval process is too slow.

The AI generates daily redistribution recommendations that nurses ignore. Alert fatigue sets in. The system gets turned off. $450K wasted, not because the AI failed, but because it automated a workflow that nobody actually understood.

The Clinical Analogy

You wouldn’t treat a patient without running diagnostics first. If someone comes into your ED with chest pain, you don’t immediately operate. You run an EKG, check labs, maybe get imaging. You diagnose before you treat.

When diagnostics reveal a patient isn’t healthy enough for surgery, you don’t cancel the surgery forever. You stabilize the patient first, then operate.

Same with AI. You need to understand what’s broken before you automate. Otherwise, you’re just automating the broken parts.

Introducing Precursor: Your AI-Native Diagnostic Tool

This is why we built Precursor.

Before you spend millions on AI that has a 95% chance of failure, you need to understand your operations deeply enough to know where AI will actually help, and where it will just magnify problems.

Precursor builds what we call a digital shadow of your hospital. Not a generic workflow diagram from a vendor playbook. A complete operational model of how YOUR hospital actually works. This includes all the workarounds, exceptions, and tribal knowledge that make things function in the real world.

How the Digital Shadow Works

Precursor doesn’t just document tasks. It captures the meaning behind your operations. Think of it as a goal-oriented operator’s manual for your hospital, one that AI can read cover to cover and understand, not just search.

Your people, your processes, and your technologies are modeled as interconnected agents, each with their own goals, constraints, and relationships. AI reads that model and can reason about it the same way a seasoned operations expert would, but faster and at a scale no human team can match.

What Precursor Actually Does

Precursor works in three stages. The digital shadow is built first. Then Team Intelligence goes to work. Then changes are implemented through small, measurable iterations.

Stage 1: Build the Digital Shadow

Your people, processes, and technologies are modeled as interconnected agents — each with their own goals, constraints, and relationships. The result is a living blueprint that AI can read, reason about, and use to ensure every decision aligns with your organizational goals. Not a snapshot. Not a diagram. A working model of how your operation actually runs.

Stage 2: Analyze with Team Intelligence

This is where clients spend most of their time with Precursor. Team Intelligence is a panel of nine AI experts — each grounded in the professional standards of their discipline — that reads your digital shadow and tells you what it sees.

Think of it as having a department’s worth of senior advisors on call — without the overhead. Each expert brings deep, standards-based knowledge from their field. They don’t start from a blank page. They start from best practice.

The nine Team Intelligence experts and what they bring:

• Team Leader — Strategic oversight and decision support, grounded in executive leadership standards.

• Business Analyst — Requirements, ROI, and gap analysis, drawing from the IIBA body of knowledge.

• Management Engineer — Process optimization grounded in IISE/IIE industrial engineering standards.

• Project Manager — Execution, timelines, and risk, drawing from PMI/PMBOK standards.

• Product Manager — Roadmap, prioritization, and feature alignment using product management standards.

• Training Coordinator — Learning design and competency development, grounded in ATD/SHRM standards.

• Systems Architect — Technical design and integration, drawing from IEEE and TOGAF standards.

• Systems Engineer — Implementation and system behavior, grounded in the INCOSE Systems Engineering Handbook.

• Management Advisor — Change leadership and adoption, grounded in change management standards.

These experts analyze your digital shadow together — surfacing misalignments, flagging where AI is likely to cause problems, and identifying where it will actually move your goals forward. They also model proposed changes before anything is implemented, so you can see predicted outcomes rather than hoping for the best.

Stage 3: Implement, Observe, and Learn

Changes go in small, deliberately. Each iteration is observable and measurable. Predicted outcomes are compared to actual results. The digital shadow updates. Team Intelligence refines its recommendations. Over time, your operation and your AI investments align — and stay aligned as things change.

Precursor in Action:

Before deploying AI, Precursor’s digital shadow would have asked five critical questions about the IV pump workflow:

1. How does equipment actually move? Digital shadow reveals: 60% of movement happens outside formal checkout process.

2. Why do staff avoid the formal process? Tribal knowledge capture reveals: 8-minute checkout process vs. 30-second workaround. Staff choose speed.

3. What goals does the workaround serve? Behavioral analysis shows: Staff prioritize patient care over documentation. The workaround isn’t laziness — it’s rational response to broken process.

4. What happens if we automate the current state? Digital shadow simulation predicts: AI will optimize based on 40% of data, missing 60% of actual equipment movement. Recommendations will be ignored.

5. What needs to change before AI can help? Gap analysis reveals: Fix checkout process first. Make formal workflow faster than the workaround. THEN add AI.

The Precursor Difference: The AI wasn't the problem. The workflow was. Precursor found it before a single dollar was spent on deployment. That's not just ROI. That's operational intelligence that lasts.

The Question Every Hospital Should Ask

Precursor is how you get certainty. It’s the diagnostic that reveals whether your operations are healthy enough for new technology. It’s the pre-flight check that prevents expensive failures. It’s the digital shadow that makes AI specific to YOUR hospital instead of generic to ‘hospitals in general.’

The question isn't whether AI will transform healthcare. The question is whether your operation is ready for it.

Visit precursoragentic.com to learn more or schedule a conversation.

Until next week,

Bryan Small ​Location Based Services Consulting ​​​