What Every Board Should Ask Before Funding AI

Listen on

Episode Description

AI is everywhere — automating customer service, optimising incredibly complex supply chains. The buzz and the possibilities seem endless, with huge amounts of excitement and money flowing in.

But here’s the uncomfortable truth: despite all that enthusiasm, about 88% of AI pilots never make it to production. They fail quietly, burning through budgets and eroding confidence.

The Real Culprit: Foundational Readiness

The problem is rarely the AI model itself. More often, it’s a lack of foundational readiness — like trying to build a skyscraper on swampy ground.

That lack of readiness can mean:

• Poor data quality
• Fragmented systems that can’t talk to each other
• Murky ownership of data or processes
• No operational framework to scale AI

For leadership, this is not a minor IT hiccup. It’s a governance issue and an investment oversight problem.

Why This Matters

In this deep dive, we’re exploring why AI ambition so often outpaces capability, and what specific actions are essential for real AI success. Whether you’re directly involved in AI projects or simply want to stay ahead of where technology is heading, understanding these systemic issues is vital.

Case Study: A European Manufacturer

A large European manufacturer was eager to jump on the AI bandwagon. They launched pilots for visual inspection, predictive maintenance, and demand forecasting.

The technology worked in isolation. But scaling across the business? Nothing stuck.

An internal AI readiness assessment revealed a low score of 2.4 out of 5. The root causes: incomplete data pipelines, ad hoc governance processes, and inconsistent integration with core systems. The board had approved significant AI spending without fully grasping these gaps.

The Turning Point

After these failures, leadership commissioned a comprehensive AI readiness assessment.

It identified where to invest first:

Stronger data governance

Robust MLOps capabilities (machine learning operations for deploying and maintaining models)

They addressed these gaps before relaunching the pilots. The result: €6 million in business value within 18 months.

The Board’s Pivotal Role

Boards are not just there to approve budgets — they set the tone for how emerging technologies are adopted. Their role is to safeguard value by ensuring investments are strategic and risks are mitigated.

For AI, that means asking: Does the foundation exist to scale successfully? Without this oversight, AI projects risk becoming expensive experiments that never deliver shareholder value.

Operationalising Oversight

The guide we reference outlines specific steps boards should take:

1. Make AI Readiness Assessments Mandatory
Evaluate data quality, system integration, AI governance, skills, and MLOps processes before approving any pilot.

2. Link AI Initiatives to Strategic Goals
Every pilot should directly advance measurable business objectives, not just serve as a technical proof of concept.

3. Govern Data as a Strategic Asset
Boards must hold management accountable for data governance covering quality, privacy, security, and ownership.

4. Monitor Readiness Continuously
Readiness is dynamic. Boards should track key indicators like clean data percentage, integration coverage, number of models in production, time-to-value, and compliance status.

5. Sequence Investments Intelligently
Fund foundational capabilities first, then move to advanced AI initiatives.

Final Takeaway

AI pilot failures are more than missed opportunities — they’re signs that governance hasn’t caught up with ambition. The most successful organisations treat AI readiness as a prerequisite, not an afterthought.

Before approving any AI budget, boards should know their readiness score, understand the gaps, and take targeted action to build a foundation that can scale AI reliably. The organisations that win in the AI era won’t necessarily be those that start fastest, but those whose leadership ensures they are truly prepared.