UNCLASSIFIED

Course 4.5: Advanced Workshop (Locked Tenant Reality)

Expert-Driven Development · 4-Hour Workshop · SSgt Jesse C. Morgan, MCD-Monterey
Name: Unit: Date:

Module Deliverables (Fill In During Class)

1. Module 1 Frontier Map Row

Pick one task in your daily workflow. Map it: AI capability column, your unit's stack, frontier issue, mitigation.
Task name:
AI capability (genai.mil / Ask Sage / both):
Frontier vs Context vs Meta-frontier:
One sentence on mitigation:

2. Module 2 Readiness Dashboard Build Notes

Record where you switched modes, where you switched tools, and where the AI failed.
Phase you switched to Cyborg:
Phase you switched between genai.mil and Ask Sage:
One concrete AI failure (be specific):
Time-with-AI vs time-without estimate:

3. Module 3 Debugging Pattern

Document one bug from group debugging. Real or canned.
Bug name (one line):
Frontier classification:
Root cause:
Root cause (cont.):
Pattern you'll remember:

4. Module 4 QA Findings

Document one planted-error you caught AND one you missed (be honest).
Caught:
Missed:
The verification habit you're committing to:
 

5. Module 5 Teach-Back Concept

Which concept did you teach? Who did you teach? What was the hardest question?
Concept:
Audience (rank, billet):
Hardest question:
How you'd handle it next time:

6. Module 6 Workflow Playbook Seed

One recurring task you'll automate this month. Pick something concrete.
Task + frequency:
Mode (Centaur / Cyborg):
Which tool (genai.mil / Ask Sage / both):
4-step outline:
Time savings goal per week:

7. Capability Gap Map Entries (NEW for 4.5)

Two rows for the Capability Gap Map. What would the Production layer unlock? Name the specific connector, license, admin grant, or policy that gates each one.
Gap 1 (workflow to production layer wish):
Gap 1 (cont.):
Gap 2 (workflow to production layer wish):
Gap 2 (cont.):

8. Assessment Self-Score

Score yourself on the six rubric criteria. Honest self-assessment. Threshold for course completion is 5 of 6 at "Meets" or better.
Frontier Map Completeness:
Complex Build Quality:
Debugging Contribution:
QA Protocol Rigor:
Teaching Effectiveness:
Workflow Playbook Completeness:
Met threshold (5 of 6 Meets)? yes / no:

EDD Framework Reminders

Centaur: distinct phases. Human designs. AI builds. Human verifies. Human refines. Human accepts. Best for strict business rules.

Cyborg: continuous back-and-forth. Discover requirements while building. Best for exploratory work where the shape is not yet clear.

Tool choice: genai.mil and Ask Sage are both first-class. Pick per task. Note which one handled the work and why. Switch when one stalls.

Frontier Recognition: name the specific task types where AI is reliable and where it is not. Capture failures in real time. Your frontier map gets sharper every cycle.

Capability Gap Map: name the wall. Specific connector. Specific license. Specific permission. Beneficiary count and time savings. That row becomes policy evidence.

Public site: jeranaias.github.io/ExpertDrivenDevelopment · Repo: github.com/jeranaias/ExpertDrivenDevelopment
UNCLASSIFIED

Power Apps skills die when you turn in your CAC. HTML skills don't. Knowing when to switch tools makes both skills last.