Define the Briefing Decision and V0 Boundary

What you will complete

Complete the Agent Charter with one decision the brief should improve, one audience, one topic band, and prohibited actions.

Decision rule: A useful intelligence agent improves a named review decision. It does not promise complete coverage or prediction.

Start with the decision, not the feed

A list of links is not yet intelligence. Name the person who reads the brief and the decision they face that morning. The purpose might be preparing client talking points, noticing product changes, reviewing policy developments, or choosing one subject for deeper research. Keep one dominant purpose for the first version so ranking rules do not compete.

Constrain the first operating loop

Set one topic band, one lookback, one run time, one report destination, and a maximum number of findings. The supplied Morning Brew guide uses five to ten sources and one report destination. The Trend Watcher guide groups sources into three to five lanes and ranks five useful developments. These are scope controls, not proof that five findings represent the whole market.

Write the human boundary before collection

The agent may collect public records, preserve evidence, normalize fields, apply approved rules, and prepare recommendations. A person decides whether a finding is reliable enough to use, whether a client should hear it, and whether any external action is appropriate. Publishing, outreach, purchasing, and consequential business changes remain outside this course.

Worked example: Northstar Workflow Studio

Fictional teaching fixture: Northstar Workflow Studio is a fictional Toronto automation consultancy. Its operator wants a weekday 7:00 a.m. brief about verified changes in AI workflow software, agent reliability, and Canadian AI policy. The report should improve one decision: which developments deserve attention in client conversations today. It will not draft or send posts, contact vendors, or claim to forecast outcomes automatically.

Northstar is not a Flowgrammer client, and its sources, weights, records, scores, timing, and results are not live evidence. The example exists so you can inspect how one choice changes the same capstone from lesson to lesson.

Workbook exercise

Open section 1 of the Morning Intelligence Agent Launch Pack. Complete the following work for your own approved topic:

1. Name the reader and the decision the report should improve.
2. Choose one topic band, lookback, morning run time, and destination.
3. Set a maximum of five primary developments.
4. List prohibited actions and the person who may approve later changes.

Write assumptions as assumptions. Keep any untested field marked untested. Ask another person to challenge one rule before you continue.

Common failure modes

• Trying to serve several audiences with conflicting priorities.
• Treating daily collection as complete market coverage.
• Combining research, publishing, outreach, and dashboards in the first release.
• Using a vague goal such as stay informed without a review decision.

Repair the smallest responsible layer: scope, source policy, collection control, evidence model, ranking rule, report, or operating gate. Do not hide a failed condition inside a more confident summary.

Capstone connection

Complete the Agent Charter with one decision the brief should improve, one audience, one topic band, and prohibited actions. This section must be specific enough for another operator to inspect, test, and reject. A complete worksheet is not automatically a passing worksheet; the evidence and critical gates still control the decision.

Check your application