Narrative Style

• Blend story-driven intros (incidents, reorganizations, migrations) with prescriptive playbooks.
• Anchor advice in measurable signals: MTTA/MTTR, defect escape rate, hiring throughput, on-call load, compliance exceptions per quarter.
• Whenever possible, show before/after org charts, RACI tables, or scorecards.

Topics to Prioritize

• Building AI-augmented teams (Cursor workflows, code review assist, rule harmonization).
• Compliance-aware leadership: how to operationalize SOX, SOC 2, Reg E obligations inside engineering processes.
• Systems thinking: treat incentives, process, and tooling as one system; highlight failure loops and feedback mechanisms.
• Incident response: runbooks, tabletop exercises, real comms templates, and post-incident governance.

Guidance Structure

1. Context – company stage, team size, regulatory burden.
2. Symptoms – metrics or anecdotes showing the pain.
3. Playbook – 30/60/90 or phased checklist with owners.
4. Instrumentation – dashboards, leading indicators, cultural signals.
5. Follow-through – cadences for retros, coaching, and audits.

Tone & Boundaries

• Avoid generic platitudes (“communicate better”); instead, prescribe mechanisms (daily risk review, RFC templates, AI pair-rotation).
• Emphasize empathy but keep language direct and actionable.
• Tie leadership moves back to business outcomes (reduced fraud exposure, fewer ACH returns, faster release cadence).
• When referencing AI tooling, mention .cursorrules, cursor/rules, CI proof, and cross-team onboarding templates.

Do / Don’t

• Do quantify cultural or operational change (e.g., “MTTR dropped from 110 → 35 minutes after adopting the drill-down template”).
• Do provide artifact links or filenames (assets/templates/team-onboarding.md, .agent/skills/*.md) when suggesting process updates.
• Do contrast old vs. new workflows with tables or checklists; readers should see exactly what changes.
• Don’t rely on inspirational quotes, leadership cliches, or anonymous “VP said…” stories.
• Don’t propose structural changes without owners, cadences, or success metrics.
• Don’t describe AI workflows without referencing validation (tests, CI scripts, rule packs).