Why AI ROI is falling short (and what leading organisations are doing differently)

Why aren’t we seeing the value yet from AI?
This is the question many organisations are asking after 18 months of pilots, experimentation and proof-of-concepts.  

For many leaders, the promised productivity gains, operational efficiencies and transformation aren’t materialising as expected. 

Anya FitzGibbon, Cuscal’s Head of Data & AI, says the problem often isn’t the technology itself. It’s how organisations are approaching it. 

In Anya’s experience, AI is treated like a technology rollout rather than what it is — an organisational transformation.  One that requires changes in behaviour, leadership, governance and the way people work. 

“The organisations seeing value from AI aren’t necessarily the ones with the most advanced technology,” says Anya. “They’re the ones helping all their people understand it, trust it and use it confidently.” 

“Technology creates the opportunity. People create the value.” 

Drawing on her experience leading AI adoption at Cuscal, Anya shares the three things the organisations seeing value from AI are doing differently. 

“You’ll start seeing the benefits when everyone touches AI” 

One of the biggest misconceptions organisations have about AI is what creates value. 

Many are still searching for one breakthrough use case or transformational project. But according to Anya, the organisations seeing the greatest returns are the ones embedding AI into everyday work across the business. 

Rather than coming from a single use case, value is often created through hundreds of small improvements that compound across the organisation, saving time, strengthening decisions and helping people work more effectively. 

“You’ll start seeing the actual benefits from AI when everyone touches it. If everyone improves what they do by 10%, that’s huge.” 

For that to happen, though, people first need to understand what AI can do. Employees can’t identify opportunities if they don’t understand the technology or how it might apply to their work. 

That’s why Anya has focused heavily on “uplifting AI literacy” across the whole business to improve uptake at Cuscal. 

“We’re not all Olympic athletes… But everyone now can do a push-up.” 

The focus hasn’t been on creating AI experts. Instead, the goal has been to build enough understanding and confidence that employees feel comfortable experimenting with the technology and identifying opportunities to improve the way they work. 

Additionally, rather than focusing on employees already enthusiastic about AI, Anya has deliberately focused on the people least likely to use it. 

“The people who love AI will play with it regardless,” she says. “I’ve found that the biggest opportunity is with the people who are unsure, intimidated or just don’t know where to start.” 

At Cuscal, that has meant creating low-pressure opportunities for employees to experiment and learn through informal Copilot sessions, prompt sharing, and practical use cases. 

Anya says that once people begin understanding what AI can do, the learning process often becomes self-perpetuating. 

“They start coming back with ideas and that’s when you start seeing real value,” says Anya.

“People can’t be afraid to use it” 

For organisations to realise value from AI, employees need to feel confident using it. 

Confidence comes from creating environments where people can experiment safely, understand the guardrails and know what responsible AI use looks like. Without it, employees are likely to either avoid AI altogether or use unapproved tools, or ‘shadow AI’, outside organisational visibility. 

In highly regulated industries like financial services, governance is essential. But according to Anya, governance should support adoption, not discourage it. 

One of the challenges Anya sees across the industry is organisations approaching AI governance in ways that unintentionally reinforce fear or hesitation around using the technology. Policies become focused on what employees shouldn’t do, rather than helping them understand how to use AI safely and responsibly. 

Cuscal has taken a different approach has been different. Rather than expecting employees to navigate governance on their own, governance and controls have been built directly into the tools and processes employees use every day. 

“Governance works when it’s visible and practical, not bolted on at the end.”  

For example, all Cuscal employees have access to Copilot in an environment where controls are already embedded, allowing them to learn, experiment and explore within clear guardrails. 

The objective is simple: make the safe way the easy way. When people feel confident using AI, they’re far more likely to participate and identify opportunities to create value. 

Importantly, Anya says organisations also need to recognise that employee expectations are changing. People are already using AI in their personal lives and increasingly expect access to similar tools at work. Organisations that fail to provide safe, approved ways to use AI risk falling behind not only in innovation, but in attracting and retaining talent.

“Our biggest constraint now is leadership” 

While broad adoption and strong governance are critical, Anya believes the organisations seeing the greatest value from AI also have one thing in common: leaders who are actively engaged in the transformation. 

AI is a business transformation that requires leaders to set direction, create momentum and identify where AI can create value across the organisation. 

Importantly, Anya says leaders don’t need to become AI experts. But they do need enough understanding to recognise opportunities, ask better questions and make informed decisions. 

“They need to understand it enough to understand the possibilities.” 

Leadership also means being willing to navigate uncertainty. One of the challenges organisations face is that there is no proven playbook for AI adoption. 

“There isn’t a proven playbook. There isn’t a proven architecture. There isn’t a proven risk assessment.” 

The technology is evolving rapidly, use cases are still emerging and many leaders are learning alongside their teams. But the organisations seeing value aren’t waiting for certainty before acting, they’re experimenting, learning and adapting as the technology evolves. 

AI ROI won’t come from technology alone 

For organisations still struggling to realise value from AI, the challenge isn’t about the technology but creating the right conditions for people to use it effectively. 

That means building literacy across the business, embedding governance into everyday workflows and ensuring leaders are actively driving the transformation. 

While the technology will continue evolving rapidly, Anya believes one thing is becoming increasingly true: the organisations that succeed will be the ones helping their people adapt fastest.  

Afterall, technology creates the opportunity, but people create the value.  

Want to dive deeper? 

Creating value from AI requires more than access to the technology. It also depends on strong leadership, quality data and a clear approach to adoption. 

Explore more AI insights from Cuscal: 

• Why smart AI starts with smarter data  
  AI is only as effective as the data that powers it. Explore why strong data foundations remain critical to unlocking reliable, scalable AI outcomes. 

• Why leadership holds the key to AI-driven growth 
  Learn why leadership engagement is increasingly becoming the differentiator between organisations that experiment with AI and those that create lasting value from it. 

• How to build your AI playbook  
  Discover practical steps for moving beyond experimentation and developing a structured approach to AI adoption across your organisation.