90-Day AI Automation Portfolio Workbook
A printable workbook for inventorying, ranking, owning, measuring, sequencing, governing, and deciding an AI automation portfolio.
Download the portfolio workbook PDF
Use this workbook with Run a 90-Day AI Automation Portfolio.
Included
- Portfolio charter and opportunity register
- Workflow decision profiles
- Value, readiness, risk, dependency, and sensitivity review
- Ownership and human decision rights
- Baselines, targets, and acceptance cases
- 30/60/90 staging and capacity plan
- Governance, vendor evidence, and decision log
- Capstone rubric and critical gates
How to use it
Complete one section after each lesson. Start with real workflow evidence, keep unknown values visible, and record the owner and source behind every portfolio decision. Preserve rejected options and failed gates instead of rewriting the history after leadership chooses a direction. Review the finished pack with the executive sponsor, business owners, system owners, and affected operators before funding a build.
Before leadership approves the roadmap
Confirm that every active initiative has a business owner, system owner, baseline, next gate, and explicit human decision boundary. Check that dependencies fit the available 30-, 60-, and 90-day capacity. Record why any higher-scoring item waits and who can reopen that decision. A weighted score cannot override a failed ownership, evidence, safety, or acceptance gate.
Harbourline and all sample weights, scores, metrics, timing, and results are fictional learner choices. The workbook supports leadership decisions; it does not authorize implementation or prove outcomes.
Start with a portfolio you can actually compare
The nine-page PDF is a companion for planning, not a requirement to approve every idea. Begin with three candidate workflows. For each, write the business owner, the repeated task, the current evidence of cost or delay and the decision a pilot would help you make. Mark missing information explicitly.
As a practice example, compare enquiry routing, invoice review and a daily research brief. They may all consume time, but they involve different data, approval requirements and failure consequences. A high estimated benefit does not make an opportunity ready if nobody can supply representative inputs or accept the result.
Turn the ranking into a next decision
Choose one candidate and write what must be true before a pilot starts: access, a baseline, an accountable owner and examples of acceptable and unacceptable outputs. Give unresolved dependencies an owner and a date for a decision. Keep the remaining ideas visible without treating them as promised delivery.
Review the portfolio when evidence changes. A pilot that reveals difficult exceptions may need a narrower scope; a lower-ranked workflow may become practical when its inputs improve. Carry the workbook into the 90-day roadmap course, where the lesson sequence helps turn the choices into a staged plan.