EODI Diagnostic Prototype

The Executive Operating Discipline Index (EODI) is a Streamlit prototype for assessing whether an executive team has the operating discipline required to convert AI capability into enterprise value.

The core thesis is constraint-based: AI value is limited by the weakest elements of the operating model. A company can have strong AI tools, but if decision velocity, incentives, talent mobility, capital reallocation, or executive fluency are weak, AI productivity will not reliably convert into economic outcomes.

Dimensions

The app scores five EODI dimensions:

• Decision Velocity
• Incentive Alignment
• Talent Redeployment
• Capital Reallocation
• Executive Fluency

Each dimension includes four diagnostic questions scored on a 1-5 slider. The app averages each dimension, normalizes the result to 0-1, and calculates:

• EODI Structural Integrity Score using the geometric mean
• Average Maturity Score using a simple average
• Constraint Stack showing the lowest three dimensions
• Top Constraint
• Value Conversion Risk
• Executive summary
• Recommended actions
• 90-minute workshop agenda

How the Files Work Together

• app.py contains the Streamlit user interface. It collects context, captures question-level slider responses, and displays diagnostic outputs.
• questions.py contains the editable EODI dimension names, question wording, helper text, scale labels, and reverse-scoring flags. Start here when changing question wording or labels.
• skills.py contains deterministic scoring, constraint diagnosis, recommendations, executive summary, and workshop agenda logic.
• prompts.py stores reusable text instructions for a future AI-enabled version of the prototype. The current app does not call an external AI API.
• AGENTS.md gives Codex repo-level guidance for keeping the prototype simple, readable, deterministic, and executive-ready.
• requirements.txt lists the Python package needed to run the Streamlit app.

Run in Replit

1. Create a new Replit Python project.
2. Upload or import this repository.
3. Confirm requirements.txt includes:

streamlit>=1.33,<2

4. In the Replit shell, install dependencies:

pip install -r requirements.txt

5. Run the app:

streamlit run app.py --server.address=0.0.0.0 --server.port=3000

6. Open the Replit web preview.

Current Limitations

• The prototype uses deterministic scoring logic only.
• It does not call external APIs.
• Scores are directional and should be paired with stakeholder discussion.
• Recommended actions and agendas are generated from fixed templates.
• Friction-oriented questions are reverse-scored so higher final scores always indicate stronger operating discipline.