2026 AI Strategy & Innovation Framework

Six phases from AI activity to five decisions you will actually fund.

Design thinking supplies the divergence. RICE, MoSCoW and Eisenhower supply the discipline. The 25/5 rule supplies the nerve. This is the instrument that turns a crowded AI portfolio into a short, defensible list — and everything else into an explicit avoid list.

Decision Velocity = (Decision Quality × Decisions Executed) ÷ Cycle Time

Velocity is not speed. It is speed with direction — better decisions, made sooner, with the evidence attached.

The framework

Diverge like a designer. Converge like a portfolio manager. Commit like an owner.

Six phases, each with one instrument and one output. Select a phase to see how it runs.

Governance spine

Data readiness — can we evidence it? Decision accountability — who owns the call? Model oversight — how is it checked? Value verification — how do we know it worked?

Phases 3 – 5 · live

Score, sort, and cut your own portfolio

Enter your candidate AI initiatives. Decision RICE ranks them, MoSCoW and Eisenhower qualify them, and the 25/5 rule commits you. Everything below recalculates as you type.

#	Initiativethe AI intervention	Decision it servesphase 1 output	Reachdecisions / qtr	Impact0.25 – 3	Conf.evidence %	Effortperson-mo.	MoSCoWobligation	Urgencydecision window	ScoreRICE	Cut25 / 5

The Five

Obligations (Must) are seated first; the remaining slots go to the highest-scoring discretionary work. These get funded, staffed, and governed — nothing else does.

The Avoid List

Ranks 6 – 25. Not a backlog. Nothing here gets capacity, a pilot, or a steering-committee slot until one of the Five is done.

Decision RICE ranking

Score ranks. It does not decide — a high-reach, low-impact initiative can top this chart and still fail the sort.

Eisenhower placement

Strategic importance against decision window. Bubble size is the RICE score; teal rings are the Five.

The Five Scored, not selected Won't — this year

Sequenced roadmap

The Five, ordered by decision window and effort. Each horizon must produce evidence before the next is funded.

Where it comes from

Four seats, one instrument

The framework is a composite of methods that each solve one part of the problem — and fail alone.

DESIGN

Human-centered design

From leading design at a global consultancy and an enterprise design studio: start with the people making the decision, not the technology.

PRIORITIZATION

RICE · MoSCoW · Eisenhower

The scoring and sorting methods used for evidence-based enterprise prioritization.

COMMITMENT

The 25/5 rule

Buffett's discipline, applied to AI portfolios: list 25, circle 5, and treat the other 20 as things to avoid until the 5 are done.

DELIVERY

Enterprise reality

Thirty-five years across Fortune 500, global consultancies, and enterprise software — where the roadmap meets the operating model.