AI executive brief: FACTS to DECISION—not yesterday's slide deck

Friday night: a PM builds the board deck—green charts, optimistic timeline, tiny risk footnote. Monday before the meeting, the CTO discovers CI pipelines are red and two epics are overdue without owners—but the slide deck already went out.

An AI executive brief earns trust when it comes from live data. In WKFGo, the MCP tool executive_brief accepts a role parameter (ceo, cfo, cto, coo, chro, cpo, ciso) and returns structured output: FACTS from tasks, git, and finance; DIAGNOSIS with ranked hypotheses; proposed DECISIONS—grounded in PostgreSQL, not PM narrative.

This is not a replacement for judgment. It is a replacement for assembling fiction from memory at midnight.

Symptoms and cost of broken executive briefs

Good briefs are decision-ready: evidence, tradeoffs, timing, owners. Bad briefs are morale theater—pretty charts that age poorly by Monday.

The hidden cost is strategic delay. When leadership spends the meeting reconciling numbers instead of choosing among options, the organization pays in slipped releases, unowned risks, and repeated debates about the same overdue epics.

executive_brief output structure in WKFGo

The executive_brief tool applies role-specific lenses:

Output includes diagnosis (hypotheses with confidence), trend (did the last decision move its target metric?), and for CEO board.tensions where officers disagree and debate is required. Honest dataGaps appear when finance or git data is incomplete—partial truth beats confident invention.

Six-step weekly brief framework

Step 1: Pull raw brief one day ahead

In Claude or Cursor: executive_brief role=cto projectId=Atlas (repeat for CFO if needed). Store output in a wiki page titled "Brief 2026-W30" so the team shares one artifact.

Step 2: PM spot-checks FACTS

Verify three random facts in the web app: overdue count on /dashboard, pending approvals on /task-approvals, git pipeline status in Git integration. On mismatch, fix data—not slides.

Step 3: Human-edit the DECISION section

AI proposes three decisions—PM assigns real owners and timing. Log with record_decision for audit. Decisions without owners are wishes, not commitments.

Step 4: Pair every brief with decision_inbox

Every weekly brief should reference at least one open decision_inbox item—or explicitly state "no decisions required this week." Briefs without inbox action are information only; they do not change behavior.

Step 5: Run a 30-minute exception-only meeting

Agenda from the brief—not round-robin. Debate board.tensions for CEO lens. Use simulate_scenario when scope or release tradeoffs appear.

Step 6: Post-meeting accountability

Call resolve_decision_item for decided inbox items. Next week's trend shows whether metrics moved. This closes the loop between narrative and reality.

Anti-patterns

• Copying briefs into PowerPoint without verify—WKFGo remains the source of truth.
• One role=ceo for everyone—CFO and CTO data needs differ sharply.
• Ignoring dataGaps in output—honest gaps beat invented numbers in board settings.
• Briefs without linking to decision_inbox—insights without an action queue evaporate.
• Treating internal briefs as auto investor decks—external narrative needs human polish and legal review.

WKFGo leadership ecosystem

• executive_brief — role-specific summary with FACTS, diagnosis, and decisions
• decision_inbox — prioritized action queue from background monitor or add_decision_item
• simulate_scenario — model delay, cut scope, add people, or freeze before committing in the brief
• portfolio_overview — multi-project CEO view
• finance_summary, get_project_report — deep dives when the brief surfaces a gap
• /reports — UI alternative without MCP for directors who prefer the browser

All permission-scoped via FeatureAccess—directors without finance access cannot pull confidential summaries. Personal wk_ keys attribute MCP calls per user.

Interpreting CEO board.tensions

When executive_brief role=ceo returns CFO vs CTO disagreement:

• CFO lens: Burn rate plus contract penalty risk from finance_summary
• CTO lens: Quality debt plus red pipelines from git integration

The CEO's job is to resolve—not average. Run simulate_scenario on each option. Call record_decision with rationale. Next week's trend calibrates whether the forecast was accurate. Structured debate beats political sidebar.

FAQ — AI executive brief

Does it replace a BI tool?
No—operational PM data from tasks, git, and finance, not a full data warehouse. Complementary to existing analytics.

How often should we run briefs?
Weekly for directors; daily exception scans via list_tasks and git for engineering leads; ad-hoc before release decisions via decision_inbox.

Can non-English executives use it?
Prompt in Persian or English; mixed FA/EN task data is fine—the summary language follows your prompt.

Is output 100% accurate?
No—spot-check FACTS in the web app. AI synthesizes; it does not oracle. dataGaps tell you where to dig manually.

Next step

Pull one executive_brief for your active project before next week's leadership sync. Spot-check three facts. Link it to the top decision_inbox item. That single cycle beats another hand-built deck.

• Sign up
• Features
• MCP docs
• Pricing
• Blog