The gap between AI adoption and actual business results is staggering: while 95% of organizations now have an AI strategy, fewer than 10% report established enterprise-level value. This paradox reveals a fundamental misunderstanding about what AI transformation actually requires. According to business leaders and consultants, the problem isn't that companies lack AI tools or ambition. The problem is that they're bolting AI onto organizational structures designed for a completely different era.

Why Are Companies Investing in AI But Not Seeing Results?

The disconnect between AI investment and measurable returns stems from a critical misconception: that AI is primarily a technology problem. In reality, organizations are treating AI as a collection of isolated tools rather than as a catalyst for redesigning how work actually gets done. When companies implement AI this way, individual tasks may become faster, but enterprise-wide productivity doesn't shift at the same rate.

Consider a traditional asset management workflow in utilities or infrastructure companies. Engineers manually gather information, prepare business cases, and navigate multiple approval pathways. AI might make each step faster, but the underlying question remains unanswered: why is the decision pathway designed that way in the first place? According to KPMG Partner Tammy Falconer, speaking at Ozwater'26, this represents the fundamental challenge facing enterprises today.

"The water sector is not behind on technology. Utilities have invested heavily in digital assets, analytics, sensors, automation, mobility, control systems, hydraulic models, asset systems, and data platforms. This is not a laggard industry. We have modernised the assets. We have digitised parts of our networks. We have introduced new platforms and dashboards. But while the technology has advanced, our organisations and the way we operate have not," Falconer stated.

Tammy Falconer, KPMG Partner

The productivity paradox is particularly stark when examining how AI is actually being deployed. Around 40% of organizations are scaling AI or driving adoption across the enterprise, but only 8% report established return on investment. This means that for every organization seeing measurable productivity gains, roughly four others are investing heavily without seeing comparable results.

What's Actually Blocking Enterprise AI Success?

The barriers to AI success are organizational, not technological. When companies fail to redesign their operating models around AI, several structural problems persist:

• Fragmented Workflows: Entire workflows remain disconnected, with handoffs and approval processes unchanged from the pre-AI era, preventing AI from creating end-to-end value.
• Slow Decision Rights: Decision-making authority remains hierarchical and slow, even as AI systems can generate insights and recommendations in real time.
• Data Silos: Critical information continues to sit in different systems, preventing AI from accessing the complete context needed for intelligent recommendations.
• Misaligned Roles: Job descriptions and team structures haven't been redesigned around AI capabilities, leaving humans and machines working at cross-purposes.
• Reactive Governance: Organizations add governance and compliance rules after AI is already deployed, rather than building them into the system from the start.

The emergence of AI agents capable of coordinating work across multiple systems is actually exposing these limitations more clearly. While individual AI use cases have become more accessible to deploy, orchestrating AI across entire systems, workflows, and decision environments is significantly more complex. This complexity is fundamentally organizational and human, not technological.

Falconer drew an analogy to Henry Ford's famous quote about faster horses: "If I had asked people what they wanted, they would have said faster horses." The water sector, and indeed most enterprises, have been asking how to make existing processes faster. But AI doesn't offer a faster horse. It offers a fundamentally different way of moving. The real opportunity isn't to take a slow process and make each step slightly faster. It's to ask whether the process needs to exist that way at all.

How Should Organizations Redesign Work Around AI?

Leading consulting firms are now providing frameworks to help organizations move beyond isolated AI pilots toward genuine operating model transformation. Guidehouse's Tech Guide 2026 outlines a five-part framework for operationalizing intelligence across the enterprise:

• Define the Work: Clearly articulate what problems AI should solve and what outcomes matter most to the business.
• Build for Intelligent Execution: Design systems and workflows that integrate data, AI, automation, and human judgment into a cohesive process.
• Activate People and Processes: Ensure employees understand their new roles and that processes are redesigned to support AI-augmented decision-making.
• Orchestrate Agents: Deploy AI agents that can coordinate work across multiple systems and workflows, not just optimize individual tasks.
• Measure and Scale Value: Track measurable outcomes and systematically expand successful approaches across the organization.

The distinction between the two paths organizations can take is crucial. Some companies use AI to improve existing processes through greater speed and efficiency. Others redesign work by integrating data, AI, automation, and human judgment into continuously learning operating models. While both approaches can drive results, organizations that redesign work around intelligence are best positioned to unlock long-term value and competitive advantage.

"AI adoption is no longer the differentiator. Enterprise intelligence is. AI is moving faster than any shift most leaders have managed before. In complex, regulated industries, speed matters, but speed without assurance creates new risk. The organizations that lead will treat AI not as a tool to install, but as a transformation of how work is designed, governed, and delivered," explained Stuart Brown, Chief Transformation Officer at Guidehouse.

Stuart Brown, Chief Transformation Officer at Guidehouse

Why Are Employees Already Using AI Outside Official Channels?

A striking finding from shadow AI research reveals that around 58% of employees use AI weekly, but approximately 44% admit to using AI in ways that breach company policy. This isn't a sign of recklessness or rule-breaking for its own sake. Rather, it reflects a fundamental mismatch between what employees need to do their jobs and what their official tools and processes allow.

Employees turn to unauthorized AI because official tools are too slow, systems are clunky, expertise is scarce, and they're under pressure to get work done. Shadow AI emerges where the formal operating model cannot keep up with the speed of work. Rather than simply restricting AI use, forward-thinking organizations should ask why employees are seeking it out in the first place, and then redesign their operating models to address those underlying needs.

This reality underscores a critical leadership challenge: the window for controlled AI transformation is narrowing. Employees are already experimenting with AI tools. The question for leaders is whether they will take control of that transformation or allow it to happen in the shadows, creating security and governance risks.

What Does This Mean for Specific Industries?

The stakes are particularly high in critical infrastructure sectors like water utilities, energy, and healthcare. These industries face climate variability, aging infrastructure, workforce shortages, and rising customer expectations. AI offers unprecedented opportunities to capture and reuse knowledge, support faster decisions, and extend expert guidance to remote locations. But only if organizations are prepared to redesign themselves around it.

The real opportunity for AI in these sectors isn't to replace human judgment, but to make it more available and more informed. AI can continuously monitor asset risk, cost, performance, and customer impact. It can generate investment options when decision thresholds are reached and route them to human decision-makers with evidence already prepared. That's not a faster horse. That's a fundamentally new way of moving.

Beyond traditional sectors, organizations across industries are beginning to recognize that the next competitive advantage won't come from deploying more AI tools. It will come from redesigning workflows, decisions, and operating models around intelligence. The organizations that move fastest on this transformation, while maintaining security and governance, will pull ahead of competitors still treating AI as a collection of isolated experiments.

"Many organizations have already proven that AI can improve individual tasks. The next challenge is operationalizing intelligence across the enterprise. Leaders are shifting their focus from experimentation to execution, asking how AI, data, governance, and human expertise can work together to transform outcomes at scale," noted Greg Meyer, Global Technology Leader at Guidehouse.

Greg Meyer, Global Technology Leader at Guidehouse

The path forward requires courage and clarity. It requires leaders to stop asking whether their organizations have adopted AI technology, and start asking whether they have the courage to reimagine their entire operating models around it. The technology is ready. The tools exist. What's missing is the organizational will to redesign how work is actually done.