Flowgrammer

Design the Morning Report and Human Review

Turn one normalized run into a concise brief with reconciled counts, evidence, unknowns, recommendations, and review boundaries.

What you will complete

Complete the Morning Report Contract and delivery rule for one destination.

Decision rule: The concise view may hide detail initially, but it may not sever a conclusion from its evidence or uncertainty.

Design for the first twenty seconds

Lead with run status, scope, and two to five evidence-bound observations. Then show ranked developments with what happened, why it may matter, source, confidence, and suggested review angle. Keep a visible section for ignored items, unknowns, exclusions, and errors.

Render from one normalized representation

The human-readable report and any JSON, CSV, or Markdown export should share the same run identifier, record identifiers, classifications, evidence references, and reconciliation counts. Escape untrusted text and allow only safe public links. A mismatch between the report and export is a failed run.

Choose one delivery destination

The first release may write a local Markdown file or create an approval-gated copy in one approved destination. Delivery failure must remain visible. This course does not authorize sending email, posting to Discord, writing social drafts, or publishing content. Those are later integrations with their own permissions.

Worked example: Northstar Workflow Studio

Fictional teaching fixture: Northstar's fictional report opens with a three-bullet summary, then lists four selected developments. Each detail shows observed fact, interpretation, suggested client-talking-point review, confidence, and evidence link. One blocked source appears under collection failures. The report says the canary is unrun and approval-dependent until a real capped run is authorized.

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

  1. Draft the report header and reconciliation block.
  2. Define the selected-development card and evidence view.
  3. Add unknown, duplicate, excluded, error, and ignored sections.
  4. Choose one destination and specify a visible delivery-failure state.

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

Common failure modes

  • A polished summary with no source trace.
  • Different counts in the report and export.
  • Delivery success recorded before the destination is checked.
  • A recommendation presented as a quoted fact.

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 Morning Report Contract and delivery rule for one destination. 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

1. What should every selected development expose?

2. The report shows four records and the export shows five. What is the run status?

3. What should happen when the destination write fails?

Next

Continue to Module 7 and carry forward the same launch pack.

Supporting source: AI Builder Café and the supplied Flowgrammer evidence-first Market News Monitor materials.

Course overview

Return to Build a Morning Intelligence and Trend-Watching Agent