Project office
Why this function is part of the AI Operating Model
Most large companies already have a corporate project office with its own project portfolio, budget cycle, reporting to leadership, and prioritization rules. The AI function arrives with its own business funnel and delivery track. If the two are not aligned, two problems appear:
- Double bookkeeping. The AI initiative lead maintains status in the AI Operating Model, while the PMO requires the same status in the corporate system — in a different format and on a different cadence.
- Competition for resources. AI initiatives compete for people and budget with regular projects, but are assessed by different criteria, and leadership does not see a single picture.
The interface with the project office is needed so that an AI initiative becomes a first-class object of the corporate portfolio, rather than a separate "shadow" alongside it.
Where it connects
| AI Operating Model stage | Role of the project office |
|---|---|
| New → Assessment | Registers the initiative in the corporate portfolio, assigns an identifier |
| Assessment → Delivery | Confirms budget and resources, opens the project/budget code |
| Delivery | Rolls up AI initiative statuses into the overall portfolio report |
| Awaiting impact → On support | Closes the project phase, transitions to operations mode |
What the function receives as input
- The AI initiative card and business case (expected impact, impact owner).
- The implementation plan and effort estimate from the delivery track.
- A dependency map for other projects and systems.
What the function delivers as output
- The initiative's place in the corporate portfolio and its priority relative to other projects.
- A budget code and confirmed resources.
- A unified status format that aggregates, rather than duplicates, statuses from the AI Operating Model.
Key interface artifacts
- Status mapping: a correspondence between business funnel states (
NEW … ON_SUPPORT) and corporate project stages. - Portfolio dashboard: AI initiatives as a slice of the overall portfolio.
Antipatterns
- PMO as a "second status bookkeeper." The initiative lead spends time synchronizing two systems instead of doing the work. Cured by a single source of status with automatic aggregation.
- AI initiatives outside the portfolio. Leadership does not see AI projects in the overall list and therefore underestimates the workload and impact.
- Rigid waterfall gating on top of iterative AI development. The corporate PMO imposes heavy project rituals on a pilot that needs to test a hypothesis quickly.