For years, AI sat largely with technical teams, being considered promising, powerful, and often confined to specialist use cases. The recent boom in generative AI changed that. It pulled AI into the boardroom, raised expectations fast, and triggered a surge of investment across large organisations; especially in regulated industries, where the stakes are higher and the path to scale is harder.
Over the past year, many leadership teams have done the right things: explored use cases, launched pilots, experimented with GenAI, agents, and automation. Yet impact often stalls at the exact moment it should scale.
Boards and exec teams, that were open to experimentation and expecting AI to be revolutionary for their companies, are now asking the tough questions:
In our work with companies across different industries, the same pattern appears again and again:
AI doesn’t fail in the experimentation stage, it fails in production: when it meets real processes, real data, real constraints, and real accountability.
In regulated industries, this becomes even sharper: if you can’t govern it, you can’t scale it; if you can’t integrate it, you can’t monetise it.
The focus now is turning AI activity into measurable outcomes. That starts with one question: 'How can we generate value from AI now?'
This is the core of AI business transformation: shifting from pilots and pockets of productivity to measurable outcomes across the value chain and proving the AI ROI.
From our experience with organisations across industries, we learned that most AI programmes stall for one (or more) of three reasons.
Whether it’s revenue growth, margin improvement, or risk reduction.
Scaling slows when security, compliance, accountability, and regulatory requirements aren’t fully embedded.
PoCs work in isolation, then stall when pushed enterprise-wide.
Taken together, these are three chapters of the same story. Value only shows up when all three are addressed, but one usually becomes the constraint first. Let's explore each chapter in a bit more detail...
In these cases, organisations have often seen some local wins: faster drafting, quicker analysis, automated routing, reduced handling time. But the big question remains: where is the financial impact?
Gartner reports that 74% of CFOs are already seeing time savings from GenAI, but only 5% report cost reductions and 6% see revenue or profit uplift (2025, Gartner CFO Leadership Series). The message is clear: GenAI boosts task productivity quickly, but turning that freed capacity into measurable financial impact takes longer — it requires workflow redesign, deliberate redeployment of effort, and an operating model that can convert productivity into outcomes.
What it looks like when integration is missing What it looks like when AI is integrated in workflows
In regulated industries, value without trust is, at best, fragile; and, at worst, completely unusable.
Public expectations are moving in the same direction as regulators: more transparency, more control, and clearer accountability. The European Commission survey on AI at work found 84% of Europeans believe AI requires careful management to protect privacy and ensure transparency.
And regulation is no longer abstract: under the EU AI Act, non-compliance with prohibited AI practices can be subject to fines of up to €35m or 7% of global annual turnover (whichever is higher).
What it looks like when trust is the constraint What it looks like when trust is a priority
Many organisations have enough experimentation to prove possibility, but not enough foundation to prove reliability.
That’s why organisations often see 'pilot purgatory': promising demos, followed by production setbacks, security constraints, or data quality issues that block implementation.
IDC research found that 88% of observed AI POCs did not reach widescale deployment. In other words: the hard part is rarely ideation, but implementation.
What it looks like when fundamentals are the blocker What it looks like when the fundamentals are in placeDiscover why AI time savings often fail to deliver real AI business value, and how enterprises can bridge the productivity gap to achieve measurable ROI, margin, and revenue impact.
Discover why AI time savings often fail to deliver real AI business value, and how enterprises can bridge the productivity gap to achieve measurable ROI, margin, and revenue impact.
Discover why trust now matters more than features in deciding which AI initiatives will actually scale, learn the cost-of-proof framework, and find out how to build a smarter AI scaling roadmap for your enterprise.
Learn how to scale AI from proof-of-concept to production by overcoming data, platform, and security challenges to achieve reliable enterprise AI deployment.
AI is transforming business by automating and augmenting work end-to-end — not just individual tasks — and by improving decision-making, speed, and control. The organisations that see impact treat this as AI-driven business transformation, redesigning workflows so time saved converts into outcomes.