Project Catalyst

AI Transformation Program Kickoff

Program Director: C. Tyrrell (Pulaski Advisory Group)

Executive Sponsor: M. Kavanagh (ACME COO)

Kickoff Date: August 17, 2026

Program Duration: 36.5 months (Aug 2026 – Aug 2029)

Why Now? The Competitive & Regulatory Moment

Two Pressures, One Solution

Regulatory: CMS-0057-F prior-auth mandate takes effect January 2027. ACME's baseline compliance is a separate, smaller project. This Program authorizes the larger capability: production AI at scale, governed, defensible.

Competitive: UnitedHealth, Humana, Oscar Health already running production AI across claims, PA, and member engagement. Moving fast without governance is risky. Moving slowly while competitors deploy is unviable. This Program proves the two aren't in tension when governance is built in from the start.

Scope: Five Workstreams, Three Years

5 Workstreams
3 BRDs (Legs)
262 People

Total Program Investment: $99 Million

Where the Money Goes

  • Consulting Labor: $41.6M
  • Internal FTE: $19.6M
  • Cloud & AI Platform: $26.9M
  • Governance & CoE: $45.2M
  • Contingency: $9.0M

Annual Breakdown

  • Year 1: $27.72M (Foundation & Flagship)
  • Year 2: $41.58M (Expansion)
  • Year 3: $29.70M (Optimization & Sustain)

Why Governance-First AI?

The industry norm: Build fast, add governance later. Results: hallucination incidents, regulatory scrutiny, liability.

Our approach: Stand up AI CoE and governance before any model hits production. Independent Model Validation, fairness testing, human-override protocols built in from Day 1.

The payoff: Defensible AI at scale. When regulators ask "How do you manage this risk?", we show them the two-line-of-defense model and documented audit trail — not retrofit compliance.

Three Boards, One Decision-Making Framework

Executive Steering Board — Business, budget, risk
AI Governance Board — Model risk, responsible AI
Enterprise Architecture Board — Technical, security, integration

Phase Gates: No BRD proceeds past a gate without sign-off from all three boards. No model reaches production without Independent Model Validation clearance.

Phase 0: Foundation & Readiness (Aug–Nov 2026)

What Happens in Phase 0

Phase 0 Gate (Nov 6, 2026): Executive Steering Board sign-off on readiness findings and governance framework v1

Year 1: Foundation & Flagship (Aug 2026 – Sep 2027)

Investment: $27.72M | Headcount: 120 active | Key Deliverables: Governance framework, Platform Foundation, Production PA AI

How We Define Success

10 Anticipated Risks — We Have Plans for Each

Technical & Data

  • Legacy data fragmentation
  • Compute cost overrun
  • Vendor delivery slip
  • Model bias in underwriting

Organizational & External

  • Regulatory scrutiny on PA override rates
  • Frontline adoption resistance
  • Executive sponsor turnover
  • Shadow AI scope pressure

See: RAIDD Log for full risk detail, mitigation strategies, and trigger thresholds

Immediate Next Steps (This Week)

We're Building Defensible AI at Scale

Governance-first. Data-driven. Transparent. Three-year commitment, sustained excellence.

Questions? Reach out to C. Tyrrell (Program Director) or your team lead

Resources: Program Charter, RAIDD Log, and Resource Plan posted to project portal (access info in your onboarding packet)

1 / 13