Use lessons 01–01 or equivalent skills before this exercise. You can revisit them from the lesson list.
A text editor and supplied fictional decision packets. No purchase, procurement action or operational approval is performed.
Start with the idea
Ask where data goes and which actions the whole system allows.
Map the model, search, logs and tools separately. Local inference does not mean the whole system is offline. Reading and sending need different permissions.
System boundary: The components, data paths and actions included in the system being reviewed.
See the example and check your understanding
A made-up teaching example
The offering runs inference on an office computer. Embeddings use a remote service, logs use a hosted dashboard, and the agent can send messages. Those paths need separate review.
Ask yourself
Can the vendor call the entire workflow offline because inference is local?
Show the explanation
No. Ask for destination, retention, access and action-authority evidence before sensitive use.
Start with these checks
- Keep remote destinations visible.
- Separate read access from sending approval.
- Do not infer compliance from appearance or location.
Use the different practice case below and check every requirement in its full instructions.
Full learning objectives
- Distinguish model, product and harness.
- Map inference and other data destinations.
- Identify decisions requiring responsible approval.
Captions appear here during playback.
Understand the complete system boundary
Ask where data goes and which actions the whole system allows.
Video playback is unavailable. Read the descriptive transcript below, go directly to practice, or open the 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
Ask where data goes and which actions the whole system allows.
What appears on screen
Ask where data goes and which actions the whole system allows.
0:06 · The idea
Map the model, search, logs and tools separately. Local inference does not mean the whole system is offline. Reading and sending need different permissions.
What appears on screen
Map the model, search, logs and tools separately. Local inference does not mean the whole system is offline. Reading and sending need different permissions.
0:19 · One useful term
System boundary. The components, data paths and actions included in the system being reviewed.
What appears on screen
The components, data paths and actions included in the system being reviewed.
0:29 · A worked example
The offering runs inference on an office computer. Embeddings use a remote service, logs use a hosted dashboard, and the agent can send messages. Those paths need separate review.
What appears on screen
The offering runs inference on an office computer. Embeddings use a remote service, logs use a hosted dashboard, and the agent can send messages. Those paths need separate review.
0:44 · Your turn to think
Can the vendor call the entire workflow offline because inference is local? Pause and decide before continuing.
What appears on screen
Can the vendor call the entire workflow offline because inference is local? Pause the video if you need more time.
1:02 · Why this answer works
No. Ask for destination, retention, access and action-authority evidence before sensitive use.
What appears on screen
No. Ask for destination, retention, access and action-authority evidence before sensitive use.
1:12 · Check 1
Keep remote destinations visible.
What appears on screen
Keep remote destinations visible.
1:17 · Check 2
Separate read access from sending approval.
What appears on screen
Separate read access from sending approval.
1:23 · Check 3
Do not infer compliance from appearance or location.
What appears on screen
Do not infer compliance from appearance or location.
1:29 · 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 →
The method
A model’s capability, deployment location and application permissions are separate properties. A local model may still use remote embeddings, connectors or logs; a cloud product may have multiple offerings and configurations. Ask for a complete data-flow diagram covering inputs, inference, retrieval, tools, storage, telemetry, identity and operator access.
Do not infer suitability from a brand, clearance, contract label or attractive security page. Record the actual offering, configuration, authorized data categories, permitted actions and responsible officials. In training, stop at a decision packet using fictional data. Organizational use requires review against the organization’s real requirements.
Choose your model, harness and offline route
Text-only assistant
Use supplied text. Ask for a plan or draft; do not claim browsing, code execution or file access that was unavailable.
Tool-enabled application or harness
Record which tools actually run, allowed files and actions, approval boundaries and observable results. Tools are optional unless the task needs them.
Local or cloud deployment
Inspect inference, retrieval, connectors, logs and storage separately. Use only authorized environments for real data; training fixtures are fictional.
Higher-capability or frontier model
Evaluate the exact task and configuration. Capability does not supply missing evidence or grant permission. No provider ranking or guarantee is implied.
Offline alternative
Use the supplied inputs and inspect the worked example manually. Builder exercises use local deterministic Python. No paid account or live model call is required.
Record the exact product/model label when available, tools actually used, inputs, permitted actions, observed output and checks. Live accounts or APIs may have external costs; they are optional and are not used by this page.
Read the general workflow guide →Worked example
Supplied input
Fictional offering: inference on an office computer; embeddings on a remote service; logs retained by a hosted dashboard; agent can send messages.
A checked result
This is not an entirely offline system. Document the remote destinations and sending authority; request retention, access and authorization evidence before any sensitive use.
Why this works
Inference location alone does not describe the full boundary or approve its actions.
Try a different case
Independent offering: cloud model, local document store, remote search index, browser connector, manual approval before sending. Draw the boundary and list five unresolved review questions.
Download this task and checklist →Interactive practice / Fictional fixture
Map the complete boundary
Use the supplied independent case above. Check the bounded decisions, then review your written artifact against the full acceptance criteria.
Everything runs in this browser. No AI model is called and no work is uploaded. Bounded checks cover specific facts and rules. Your writing is not graded: you review it yourself. Practice never marks a lesson complete automatically.
Include cloud inference, local store, remote index, connector, identity, retention and responsible review.
Try the task, run the checks, then revise. Checks are teaching feedback, not an instructor grade.
Observed browser simulation
This trace comes from the bounded local checker. It is not a Python run, model result or production tool execution.
Keep or restore your work
Drafts are not saved automatically. Save only fictional practice data. A browser draft may be lost if storage is cleared; download a copy to keep it.
Your steps
- Map each destination and action.
- Separate known configuration from unknown retention and access.
- Name responsible review roles without inventing their approval.
Check your work
- Remote index and browser connection remain visible.
- Approval before sending is distinguished from read access.
- No claim of compliance, accreditation or authorization.
After your attempt: review notes and a common mistake
“Documents are local, so data stays local” omits model and index flows. Ask what is transmitted and retained, and what approval actually covers.
Different wording or a different justified decision can pass. Check your work against every requirement; these notes are not an automated grade.
Explain your approach
Which missing boundary detail could invalidate the proposed use?
Keep your artifact, one correction or decision, its evidence and one remaining limitation. Nothing on this page sends or uploads your work.
Check your understanding
Choose an answer to see feedback. This is a knowledge check, not instructor assessment.
Review your independent practice
Complete the supplied task before recording progress. Use the acceptance criteria and reveal the review notes after your attempt. Keep your artifact yourself; nothing is uploaded.
Not yet self-reviewed. Completion stays in this browser and does not issue a badge.
Optional deeper study
Explore the related general-course lesson → This opens the existing canonical lesson; it is optional depth, not another copy of this focused task.
Primary references
Revision 2026-10-03. Fictional examples, not observed provider results. Knowledge checks and completion records are self-study aids.