Advanced Workshop

Overview & Learning Objectives

This workshop takes you beyond building individual tools and into the practices that make an entire section more capable. By the end of this course, you will be able to:

• Map the frontier for your specific domain
• Build verification protocols for AI-generated output
• Practice debugging complex systems
• Learn to teach the 201 skills to others

Module 1: Frontier Mapping for Your Domain

Duration: 45 minutes

The frontier is the boundary between what AI handles well and where it fails. Your job in this module is to create a detailed map of that boundary for your specific role and domain. This is the most important artifact you will produce in this course.

Example: Filled Frontier Map (1st Bn, 99th Marines)

Study this completed map to understand the level of specificity expected. Each row covers a category of work, and each column captures where AI sits today relative to that category.

Blank Frontier Map Template

Use this template to build your own frontier map. Replace the placeholder rows with the actual categories of work in your role.

Exercise: Fill In Your Frontier Map

Using the blank template above, fill in your frontier map for your specific role and domain. Be as specific as possible — generic entries like “writing” are less useful than “counseling statement drafts” or “award narratives.”

Module 2: Complex Build — Multi-Component System

Duration: 60 minutes

In this module you will build a system that requires switching between centaur and cyborg modes across multiple phases. The key skill is conscious, deliberate mode-switching — choosing the right work pattern for each component based on what that component demands.

Example: Tutoring Management System (1st Bn, 99th Marines)

This system tracks tutoring sessions for the Language and Culture Program. It has three components, each requiring a different work mode:

• Data backend — Centaur mode (accuracy-critical; schema design must be correct before anything else is built)
• Input interface — Cyborg mode (iterative, fluid collaboration with rapid back-and-forth)
• Automated reporting dashboard — Centaur mode (accuracy-critical; numbers must be right)

Phase 1: Data Backend (Centaur Mode)

Why centaur mode: The data model is the foundation. If the schema is wrong, every component built on top of it will inherit the error. You define the requirements precisely, AI generates the design, and you verify before moving on.

Design a SharePoint data model for a tutoring management system. Tables needed:
Students (name, rank, unit, language, proficiency level), Tutors (name, rank,
languages, availability), Sessions (student, tutor, date, time, duration, topic,
notes), and Reports (student, period, hours completed, progress notes).
Define the relationships between tables and required columns.

Verification at the phase boundary:

• Does the output match your design requirements?
• Are the data types correct for each column?
• Do the relationships between tables make sense?
• Are there any missing columns you will need later?

Phase 2: Input Interface (Cyborg Mode)

Why cyborg mode: The input form is user-facing and benefits from rapid iteration. You work alongside AI in a fluid back-and-forth, adjusting layout, field order, and validation rules as you go.

Build a Power Apps form for logging tutoring sessions. It should: pull student
and tutor names from the existing lists, auto-fill the date and time, have a
dropdown for topic area, and a notes field. Make it mobile-friendly so tutors
can log sessions from their phones.

Verification at the phase boundary:

• Does the form pull from the correct data sources?
• Do the dropdowns reflect the actual lists from Phase 1?
• Does the auto-fill logic work correctly?
• Is the layout usable on a mobile device?

Phase 3: Automated Reporting Dashboard (Centaur Mode)

Why centaur mode: Dashboard numbers will be briefed to leadership. Incorrect metrics erode trust in the entire system. You specify exactly what metrics are needed, AI generates the dashboard, and you verify every number against source data.

Create a Power BI dashboard that shows: total tutoring hours this month by
student, tutor utilization rates, most common topic areas, and students who
haven't attended sessions in 2+ weeks. Include a filter for date range and
language.

Verification at the phase boundary:

• Do the totals match a manual count of the source data?
• Are the utilization rate calculations correct?
• Does the “2+ weeks absent” filter use the right date logic?
• Do filters work as expected?

Module 3: Verification Protocols and QA

Duration: 45 minutes

Every AI-generated output needs a verification step. This module gives you a structured QA checklist and a timed exercise to practice using it.

QA Checklist

1. Source verification — AI fabricates references. Every citation, regulation number, and URL must be independently verified.
2. Data accuracy — Numbers, dates, names, and quantities must be checked against source data.
3. Logic check — Does the reasoning hold? Are conclusions supported by the premises?
4. Format compliance — Does the output match required formats, templates, and standards?
5. Domain review — Does this pass the smell test for someone who knows this domain?

Exercise: Timed QA Review

The following document was generated by AI. It contains five planted errors. Time yourself. Use the QA checklist above to find all five issues in this document.

Module 4: Debugging Practice

Duration: 45 minutes

Debugging is a core skill for anyone building with AI. In this module you will work through three scenarios individually, diagnosing root causes and identifying fixes. Each scenario represents a common failure pattern in AI-assisted builds.

Scenario 1: Approval Flow Sends to the Wrong Person

A Power Automate flow is supposed to route purchase requests based on dollar amount. Requests under $2,500 go to the Department Head for approval. Requests $2,500 and above go to the Commanding Officer.

Request for $1,500 office supplies submitted by Cpl Torres.
Expected: Routed to Department Head (Maj Williams).
Actual: Routed to CO (Col Richardson).

Request for $2,500 equipment purchase submitted by SSgt Park.
Expected: Routed to CO (Col Richardson).
Actual: Routed to CO (Col Richardson). [Correct]

All requests at $2,500 exactly are handled correctly. Only requests BELOW
$2,500 are being sent to the wrong approver.

Diagnosis guide:

1. What is the condition that controls routing?
2. If requests at exactly $2,500 go to the CO correctly, but requests below $2,500 also go to the CO, what comparison operator would cause this?
3. What should the condition be instead?

Scenario 2: Gallery View Shows All Items Instead of Filtered

A Power App has a gallery that should display only overdue items — items whose due date has passed and whose status is not “Complete.”

The "Overdue Items" gallery shows ALL items from the task list,
including items that are marked "Complete" and items whose due date
is in the future.

Total items in list: 47
Items actually overdue: 12
Items displayed in gallery: 47

Diagnosis guide:

1. What filter formula is the gallery using? Check the Items property of the gallery.
2. Is the filter comparing dates correctly? (Common issue: comparing a date value to a text string.)
3. Is the status check using the correct column name and value? (Common issue: column display name vs. internal name.)

Scenario 3: Dashboard Numbers Don’t Match Source Data

A Power BI dashboard reports completion percentage for a training tracker. Leadership questions the numbers because they do not match a manual count.

Dashboard shows: 85% training completion rate.
Manual count shows: 67% training completion rate.

Source data: 47 Marines, 12 training events each = 564 total slots.
Dashboard appears to be counting something differently than
"completed slots / total slots."

Diagnosis guide:

1. How is the dashboard calculating “completion rate”? Is it counting rows vs. distinct values?
2. If a Marine completed 10 of 12 events, does the dashboard count that Marine as “85% complete” or as “1 complete Marine”?
3. Are there duplicate rows in the source data inflating the count?

Module 5: Teaching Others — The 201 Multiplier

Duration: 30 minutes

Individual AI capability is valuable. Organizational AI capability is transformational. This module covers how to spread 201 skills across your section — and how to do it responsibly.

The Permission Gap

Mollick’s research shows workers are already using AI but hiding it. They are worried about organizational reaction — will leadership see it as cheating? Will they get in trouble? This creates a shadow AI culture where best practices are not shared, mistakes are repeated, and no one benefits from anyone else’s learning.

The fix is explicit permission and structure. When leadership endorses AI use through a framework like EDD, people stop hiding and start sharing.

The Apprentice Problem

Entry-level job postings dropped 35% from 2023 to 2025. AI is automating the routine tasks that juniors traditionally learned on. If a junior Marine never manually writes a counseling statement because AI generates it, how do they develop the judgment to know when the AI-generated version is wrong?

This is not a reason to stop using AI. It is a reason to be deliberate about how juniors interact with it.

Protocol for Junior Marines Using AI

• Require review and explanation — Juniors must review AI output and explain WHY it is correct or incorrect. The explanation builds the judgment that unreviewed use would skip.
• Periodically work without AI — Key tasks should periodically be done from scratch to build foundational skills. A Marine who has never written a memo manually cannot evaluate an AI-generated memo.
• Use AI output as a teaching tool — Give juniors AI-generated products and have them find the problems. This builds critical evaluation skills faster than starting from blank.
• Rotate QA review — Assign juniors to the QA review step so they develop quality judgment through repeated exposure to both good and flawed output.

Exercise: The 60-Second Teach

Pick one 201 skill you have used in your own work. Write a 60-second explanation that someone new to AI could follow. Use your own real example — not a hypothetical.

Deliverable: Workflow Playbook

Your final deliverable is a one-page playbook for one AI-integrated workflow from your actual job. Study the completed example below, then create your own using the blank template.

Completed Example: Weekly Training Schedule Publication

Blank Workflow Playbook Template

Capstone Deliverable

Knowledge Check

Course Completion Checklist