Applied workflow capstone

Deliver a checked workflow and explain what it can and cannot do.

Visual lessons: 1 · 1:43 total Read estimate: 3 min Practice estimate: 4 hr Updated: 2026-10-03
In this lesson · A8 of 8All lessons in this track →

Progress for this track is saved in this browser only. No completion badge is issued. View your learning.

Start with the idea

Deliver a checked workflow and explain what it can and cannot do.

Keep the sources, instructions, results and checks together. Show failures as well as successes. Limit your conclusion to what the exercise actually demonstrated.

Handoff: Giving another person the information needed to inspect and repeat your work.

See the example and check your understanding

A made-up teaching example

A public-event draft uses two conflicting sources. The AI-assisted process is faster, but it wrongly settles the conflict. The recommendation is to revise it, with conflict cases handled manually.

Ask yourself

Does lower effort prove the workflow is ready for operational use?

Show the explanation

No. Its important quality checks must pass too. The exercise uses invented records and does not prove a real deployment outcome.

Start with these checks

  • Trace important claims to sources.
  • Include review in the effort comparison.
  • Explain failures and remaining limits.

Use the different practice case below and check every requirement in its full instructions.

Go to the practice task ↓

The video explains the main idea. Use the sources, instructions and checks below to complete the practice.

Visual lesson1:43

Captions appear here during playback.

Applied workflow capstone

Deliver a checked workflow and explain what it can and cannot do.

Download transcriptOpen video file

Source-based teaching example · Fictional data · Reviewed 2026-10-03. Microsoft Mark · locally synthesized. On a small screen, use fullscreen for diagram details or read the transcript.

Read the descriptive transcript

0:00 · What you will learn

Deliver a checked workflow and explain what it can and cannot do.

What appears on screen

Deliver a checked workflow and explain what it can and cannot do.

0:06 · The idea

Keep the sources, instructions, results and checks together. Show failures as well as successes. Limit your conclusion to what the exercise actually demonstrated.

What appears on screen

Keep the sources, instructions, results and checks together. Show failures as well as successes. Limit your conclusion to what the exercise actually demonstrated.

0:21 · One useful term

Handoff. Giving another person the information needed to inspect and repeat your work.

What appears on screen

Giving another person the information needed to inspect and repeat your work.

0:29 · A worked example

A public-event draft uses two conflicting sources. The AI-assisted process is faster, but it wrongly settles the conflict. The recommendation is to revise it, with conflict cases handled manually.

What appears on screen

A public-event draft uses two conflicting sources. The AI-assisted process is faster, but it wrongly settles the conflict. The recommendation is to revise it, with conflict cases handled manually.

0:46 · Your turn to think

Does lower effort prove the workflow is ready for operational use? Pause and decide before continuing.

What appears on screen

Does lower effort prove the workflow is ready for operational use? Pause the video if you need more time.

1:04 · Why this answer works

No. Its important quality checks must pass too. The exercise uses invented records and does not prove a real deployment outcome.

What appears on screen

No. Its important quality checks must pass too. The exercise uses invented records and does not prove a real deployment outcome.

1:15 · Check 1

Trace important claims to sources.

What appears on screen

Trace important claims to sources.

1:21 · Check 2

Include review in the effort comparison.

What appears on screen

Include review in the effort comparison.

1:26 · Check 3

Explain failures and remaining limits.

What appears on screen

Explain failures and remaining limits.

1:31 · Now try a different case

Open the lesson and try its different practice case. Use every check in the full task. Watching alone does not mark the lesson complete.

What appears on screen

Open the lesson Try the new case Check and revise The full task and its requirements are in the linked lesson.

Watching explains the example. Complete the lesson practice to check what you can do. Explore all visual lessons →

Before you start

Prerequisite: A1–A7 artifacts; revisit failed gates before integration. Offline/no-cost route: paper, calculator and the synthetic facts below. Optional tools must already be authorized. No paid API, private workplace record, account creation or external action is required.

What you will be able to do

  • Deliver authorized workflow with baseline and evaluation.
  • Observe unaided peer execution.
  • Defend value, limits, risk, and maintenance.

Part 1 Build a package another person can inspect

Capstone evidence combines intake, authorization simulation, CTSE, data/source record, workflow configuration, baseline/evaluation, SOP, change/risk register and bounded recommendation. Reproduction tests documentation: give a peer or simulated new operator only the package, record questions, revise and repeat. Do not coach silently. A defense should withstand changed assumptions, such as missing sources, changed model or revoked permission. A release recommendation follows predeclared criteria; training results are not production validation, certification or permission to act. Preserve unfavorable results rather than selecting only impressive demos.

Part 2 Worked synthetic example

Synthetic public-event workflow uses approved notice S1: Cedar workshop 14 October, 20 seats, venue pending; S2 draft says online confirmed. A draft-only package produces a brief with conflict visible. Manual baseline 15 minutes/case; candidate draft+review 13 minutes, but candidate incorrectly resolves one conflict. Recommendation: revise and limit training use to routine sources; conflict cases require manual resolution. No successful operational deployment or employment claim is warranted.

Part 3 Guided practice and feedback

Design review before building: identify scope, data path, model/product/harness/environment, expected outputs, no-go actions and test cases. Changed assumption: S1 withdrawn. Expected response: stop treating it as current authority, seek replacement and keep the brief unresolved. Changed assumption: send permission requested. Expected: separate authorization gate; drafting scope does not imply send scope.

Review and next step

Self-review is learning evidence, not independent certification or operational approval. Use the weakest rubric item to choose further practice. Any real rollout needs its own authorized environment, representative validation and responsible owner.

Substantive review: 2026-10-02. See the configuration guide and the NIST AI RMF, a voluntary framework rather than a course authorization or credential.

Lesson status:PRACTICE REVIEW REQUIRED

Review your practice