Course-aligned practice for D31-D36 Context Template Seed. This page uses the previous and next artifacts to keep the practice connected.
Part 2 | D31-D36 Context Template Seed | normal
Name the background fields that prevent AI from inventing context or flattening the situation.
A worker says: “The goal is clear, but AI still misses the background every time.”
It sounds like adding more notes. The trap is that background fields must separate stable context, current case details, and excluded information.
“Background: everything from the customer thread and internal notes.”
The weak version contains everything, customer thread, internal notes, no stable/current split.
Reusable facts that rarely change.
Details for this task only.
Sensitive or irrelevant material kept out.
Use this rule before filling the build table. The goal is a reusable context asset, not another one-off prompt.
Reusable facts that rarely change.
Details for this task only.
Sensitive or irrelevant material kept out.
A support lead needs AI to draft an incident summary, but the model lacks product version, affected customers, timeline, and current workaround.
Name the background fields that prevent AI from inventing context or flattening the situation.
Letting AI infer background facts because the template only asks for a task goal.
AI gives generic advice because it does not know timeline, customer type, channel, or prior decision. Background fields turn “good advice” into workplace-fit output.
Which background field changes the answer most?
Received: D33 task_goal_fields.md
Hands off: D35 audience_fields.md
Redline: Do not allow AI to fill background facts without evidence notes or explicit unknown markers.
Before building the artifact, decide what the template should make easier to review.
This is not a generic sample. It is filled for Day 34: Background Fields.
| No. | Background field | Evidence needed | Why it matters | Unknown risk | Ask or repair action | Evidence used | Readiness |
|---|---|---|---|---|---|---|---|
1 | Affected customer segment | Incident ticket + CRM segment tag | Changes severity and wording | AI may overstate impact | Ask support ops for affected segment before drafting | INC-1042 + CRM segment export | Ready for D35 |
The output can now match urgency, audience, and channel.
Generic retention advice
Customer segment: enterprise renewal; channel: email; timeline: decision due Friday.
Core path: complete three rows. Deep path: complete three rows. Blank required fields create export warnings.
| No. | Background field | Evidence needed | Why it matters | Unknown risk | Ask or repair action | Evidence used | Readiness |
|---|---|---|---|---|---|---|---|
| 1 | |||||||
| 2 | |||||||
| 3 |
Optional but recommended: Use this pass to find the context/template field that would fail reuse, confuse a teammate, or hide a review need before export.
Weak row: Background includes all notes and internal comments.
Repair move: Split stable/current/excluded fields.
Better row: Stable: refund policy source; current: customer order issue; excluded: account ID and internal escalation notes.
Choose the field you would be comfortable carrying into the next day.
Choose the field most likely to fail reuse or need review.
Sentence starter: The weak row is weak because ___. I will repair it by adding ___ before using AI.
This file is today’s work receipt. It should show which context/template field can be reused tomorrow, not just that you filled a table.
What this artifact proves: This file proves you can design background fields that are reusable and safe.
Weak export: Background includes all notes and internal comments.
Good export: Stable: refund policy source; current: customer order issue; excluded: account ID and internal escalation notes.
Click Generate Markdown to create background_fields.md.
Continue with a day-specific practice example.