Skip to content
Lucas BarriosApplied AI & Operational Transformation

European Enterprise Client · 5 Business Divisions

Enterprise AI Program Office Design

Operating model design for an enterprise AI Program Office governing AI initiatives across five European business divisions — covering intake process, portfolio prioritization, risk governance, board reporting, and change management.

European Enterprise Client · 5 Business Divisions8 weeksAI Governance · Organizational Design
Business divisions governed
5

Finance, Operations, HR, Customer Experience, and IT divisions.

AI initiatives in scope
47

Active and pipeline initiatives governed across all five divisions.

Time to first governance decision
12 weeks

From program office launch to first formal governance decision.

Portfolio under review
€2.4M

Annual value of AI initiatives entering the governance process.

Context

A pan-European enterprise with five autonomous business divisions was running 47 AI initiatives with no central visibility, inconsistent risk assessment, and no clear accountability when issues arose. The board had mandated an AI governance structure ahead of EU AI Act implementation deadlines.

Prior attempts at centralized governance had failed due to divisional resistance and process overhead. The question was not whether to build a Program Office — it was how to build one that divisions would actually use.

Approach

Phase 1 · Weeks 1–2

Diagnostic

Stakeholder interviews across all five divisions, inventory of 47 active AI initiatives, failure mode analysis of the previous governance attempt, and identification of 12 high-risk systems requiring immediate attention.

Phase 2 · Weeks 3–4

Operating Model Design

Program Office charter, RACI matrix across eight roles, three-track review process (fast/standard/extended based on risk tier), intake form and scoring methodology, and documented escalation paths and decision rights.

Phase 3 · Weeks 5–6

Portfolio Architecture

Standardized initiative data model, portfolio health metrics, division-level reporting templates, board quarterly review format, and incident classification and response protocol.

Phase 4 · Weeks 7–8

Change Management & Launch

Division lead onboarding workshop design, communication plan for initiative owners, 90-day adoption monitoring plan, and vendor evaluation framework for third-party AI procurement.

Program Office Simulator

The governance process designed in this engagement is reproduced here as a working simulator. Use the intake form to classify a new AI initiative, review the live portfolio of eight governed initiatives, and step through a governance decision for a high-risk system — the same three workflows that the Program Office runs in production.

New Initiative

Submit an AI initiative for Program Office review and risk classification.

1010,000
Autonomous decisions? No

Technical Depth

Decision 1

Why three review tracks instead of one

Single-track governance fails because low-risk initiatives get stuck behind high-risk ones, creating backlog and divisional resistance. Three tracks (2/6/12 weeks) preserve speed for low-risk work while ensuring rigor where it matters. The fast track alone reduced projected review wait time for minimal-risk initiatives by more than 70%.

Decision 2

Why the Program Office reports to the CDO, not the CRO

Risk-first governance creates a compliance mindset that slows value creation. CDO ownership positions the Program Office as an enabler of AI investment, not a gatekeeper — critical for divisional adoption. The previous governance attempt failed partly because it was perceived as a control function.

Decision 3

Why change management was 25% of the engagement scope

The previous governance attempt failed not because the framework was wrong but because division leads were not involved in its design. Co-creation workshops and clear WIIFM (what's in it for me) messaging for each division were the actual unlock — not the RACI or the process flow.

Reflection

Next step

Review the supporting profile.

Use CV access and LinkedIn for background, or return to selected work for more examples of structured AI thinking.