85–87% of AI projects fail — not because the technology doesn’t work, but because organisations pick the wrong projects, underestimate change management, or lack data readiness. The Decision Engine exists to fix that.

Data Driven Not Ego

The 3-Step Pipeline

The Decision Engine replaces subjective debates with a single transparent formula that weights what actually matters in AI project success:

1

Discover

Gut feel
2×2 matrix

2

Score & Decide

5-pillar scoring
AIRI DE formula

3

Prioritise

Strategic judgement
Dot voting

Step 1 – Discovery Matrix

Start With Discovery

Before scoring anything, teams brainstorm real pain points and place project ideas on a simple 2×2 Impact × Feasibility matrix. This is a 1-minute gut-feel exercise — no rubrics, no arithmetic.

Before scoring anything, teams brainstorm real pain points and place project ideas on a simple 2×2 Impact × Feasibility matrix. This is a 1-minute gut-feel exercise — no rubrics, no arithmetic.

Selection rule: Pick 2 from Do This. If fewer than 2, pull from Moonshot first, Quick Win second. Never from Drop.

Impact x Feasibility MatriX
Step 2 – The AIRI Decision Engine

One Formula, Five Pillars, Complete Clarity.

The Decision Engine replaces subjective debates with a single transparent formula that weights what actually matters in AI project success:

AIRI DECISION ENGINE SCORE

[(0.7×P4 + 0.3×P5) + P3 + 2×P1] × P2 Ethics Gate

2X

P1 · Leadership & Culture

Leadership, talent, and change capacity. Double-weighted because people are what make or break AI projects.

1X

P3 · Business Value

Strength of the business case. Important, but a great business case without org readiness still fails.

0.7 + 0.3

P4 · Data + P5 · Infra

Data dominates infrastructure. You can spin up cloud compute; you can’t conjure clean data.

Gate

P2 · Ethics & Governance

Not a score to optimise — a gate. No governance? Full stop. The formula multiplies by zero.

The Ethics Gate

Ethics Is Not Optional. It Is A Multiplier.

P2 doesn’t add to the score — it multiplies the entire result. An organisation with no ethics framework gets zero, regardless of how strong everything else is.

x0

STOP

P2 at L0 (AI Unaware). No governance framework. Do not proceed. Build the ethics foundation first.

x0.5

CONDITIONS

P2 at L1 (AI Aware). Basic awareness but no formal framework. Proceed with an ethics remediation plan. Score is halved.

x1

GO

P2 at L2+ (AI Ready or above). Adequate governance and risk framework in place. Full score applies.

See It In Action

Worked Example

A mid-sized company evaluates deploying an AI chatbot to handle tier-1 customer queries.

Calculations

1

Feasibility: (0.7 × 2.50) + (0.3 × 3.75) = 2.875

2

+ Business Value: 2.875 + 3.75 = 6.625

3

+ 2× Leadership & Culture: 6.625 + (2 x 2.50) = 11.625

4

x Ethics Gate: 11.625 × 1 = 11.625

AIRI Decision Engine Score = 11.6 / 20 = 58%

A moderate candidate. Proceed with caution on weak pillars. P4 (Data Readiness at L2) is the bottleneck — improving data access would push the score significantly higher.

Reading the Score

What The AIRI Decision Engine Tells You

The theoretical maximum is 20. Here’s how to interpret where a project lands:

0

Ethics-killed

<25%

Very low

Drop.

25-50%

Weak

Significant gaps.

50-75%

Moderate

Go! But watch out for weak pillars.

75-100%

Strong

High confidence! Prioritise and resource.

“It’s not a rejection, it’s a roadmap.”

Step 3 – Strategic Judgment

AIRI Design Engine Prioritise. The Team Decides What To Champion.

After the top 3 – 5 projects are scored and presented, every participant votes silently across three strategic lenses. No discussion, no lobbying — just stickers on worksheets mounted on the wall.

This is deliberate. The Decision Engine removes politics from evaluation. The dot vote adds back strategic intent — what leadership actually wants to invest in, knowing the readiness data.

Rules: 2 stickers per round per person. Both stickers on one project = OK (signals conviction). No tallying between rounds — anchoring on the leader biases subsequent votes. Count everything at the end.

Three Voting Rounds

IMPACT

Would this transform our organisation in 3 years?

READINESS

Can we start meaningful progress within 90 days?

COLLABORATION

Would other organisations also benefit from this?

Tiebreakers: Highest Impact count → Highest DE Score → 3-minute discussion + show of hands.

Ready to find your top 3 AI projects?

Download the complete workshop materials — runbook, A1 worksheets, and facilitator scripts — and join the AIR Framework Community!

Download Workshop Materials Join The AIRI Framework Community