Frontier Map — Seed Worksheet

Frontier Map Seed Worksheet (Markdown)

The starter your section copies into a Teams shared note (or wiki page, or OneNote tab) at the end of Week 1 and keeps updating. Pre-formatted with the four-column map, the failure-case capture rows for Week 3, and a quarterly re-test schedule.

The four-column frontier map

Each row is one specific task type your team performs. “Writing” is too generic. “Drafting NAM award narratives” is a row. “Interpreting MCO 1000.6 for a check-in question” is a row.

• Task type. Name a task someone on the team actually performs by name. The more specific, the more useful the row is to the next person who reads the map.
• Handles well today. Drafting, summarising, restructuring, brainstorming, formatting, explaining, translating between styles — the green column from the deck.
• Handles poorly today. Specific MCO interpretation, TIS/TIG calculations, current personnel data, anything classified, anything that depends on institutional knowledge that is not in the model's training set — the red column.
• Moving frontier. The third column the deck adds verbally. Tasks where capability is improving and you should re-test on a calendar. Re-test what AI couldn't do six months ago.
• Failure cases observed. The fourth column added in Week 3. Date · who hit the failure · what went wrong. This is the column that grows fastest once builds start.

Why we capture failures: the Microsoft research cited in the Week 1 deck says it takes ~11 weeks to build the AI habit. Failure cases shorten that curve for everyone in your section. Sharing what AI got wrong saves your peers an hour of confusion.

The five-minute seed exercise

Run a 5-minute timer. While students write, you open the Teams shared note titled “[Unit] Frontier Map — Seed” in the channel. After the timer:

1. Three volunteers each share one example from each column.
2. You transcribe what they say into the shared note in real time.
3. The room adds to the master list in chat as people speak.
4. Star the failure cases that surprised the room — those are the most valuable rows.

That seeded note is what students take with them. It's the section's starter map. They'll keep adding to it through Course 2 and Course 3.

Carry-forward into Week 3 builds

Week 3 (Platform Training) revisits the map twice. Slide 15 cues the instructor to capture failure cases on a whiteboard or shared doc grouped by type — “AI invented a column”, “AI swapped two rules”, “AI's syntax was off”. Slide 27 cues students to add at least one new row to the unit's frontier map, and Slide 28 tells the instructor where to log failure cases — this shared frontier-map doc.

The fourth column on the seed worksheet exists for exactly that handoff. Builders log a failure, a verification step that would have caught it, and the row moves the unit's whole map forward.