Clean Data
Before anything goes to market, the fuel has to be right. Fix the data first and AI performs. Skip it and it fails — no matter how good your prompts are.
Not started
Mission Objective
Before anything goes to market, the fuel has to be right. Fix the data first and AI performs. Skip it and it fails — no matter how good your prompts are.
- Explain why 95% of AI programs fail and what the AI Officer does differently
- Identify the three layers of AI and why the data layer is where results are made or broken
- Clean a messy dataset and verify AI's work before trusting it
- Evaluate AI tools based on the four things that actually matter
- Make governance decisions about data safety before anything gets pasted
Mission Challenges
C1
Active
Garbage In, Garbage Out
Take the messy BOLT pilot taste test data (BOLT_Pilot_Feedback_Messy.csv) and feed it directly to your AI tool without any cleaning. Ask for an analysis.
C2
Active
Become a Data Detective
Go through the BOLT pilot data row by row. Find every problem. Do not use AI to find them - do this manually so you develop the instinct.
★ Final Project
Open
Clean at Scale
The full 100-response BOLT survey just came in (BOLT_Full_Survey_Messy.csv). Dana is presenting to leadership in 48 hours. She needs a clean dataset and a summary analysis she can trust. Your job is t
60 minutes
Begin →
Mission Briefs
Brief 1
Why Data Breaks AI
20 min
Brief 2
The Data Audit
20 min
Brief 3
Clean at Scale
25 min
