Boards are increasingly asking a simple question: where is the return on AI investment?
It’s a fair challenge. Investment is rising, expectations are high, and scrutiny is increasing. But for many organisations, that return is proving difficult to demonstrate, not simply because it is hard to evidence, but because in many cases, it is not yet being meaningfully realised.
Research reflects this. McKinsey & Company reports that while AI adoption continues to grow, only a minority of organisations are seeing material bottom-line impact. Deloitte highlights the difficulty organisations face in defining and measuring return, particularly in the context of generative AI. Boston Consulting Group goes further, showing that successful transformation is driven less by technology itself and more by people and behavioural change, areas that remain the least measured and least evidenced.
AI isn’t consistently delivering ROI
The assumption is often that deploying AI will, by default, generate value. In practice, the picture is more mixed. In many cases, AI is not yet delivering meaningful return. And where it is, organisations often lack clarity on what that return should look like, and how it should be evidenced through measurable changes in performance and outcomes. For some organisations, ROI may be reducing complaints. For others, improving first contact resolution, accelerating speed to competence, or ensuring employees apply knowledge consistently and compliantly.
The challenge is not a lack of metrics. It is a lack of clarity, alignment, and the ability to connect investment to measurable change in day-to-day performance. This is where many AI solutions fall short. They can be implemented, adopted, and even widely used, but they do not always provide a clear line of sight between usage and outcome.
Where organisations are starting to shift
What we’re seeing at Elephants Don’t Forget is a shift in how organisations are approaching this. Senior leaders are stepping back from the assumption that technology alone will deliver return, and refocusing on where value is actually realised – through their people.
This is particularly true in environments where interactions are becoming more complex, more sensitive, and higher risk. In these moments, performance is not driven by access to information alone, but by how effectively individuals apply knowledge in role. If that is where value is realised, then AI investment, and measurement, needs to follow. In many cases, it still does not.
Where most AI solutions fall short
AI investment remains largely focused on providing access to information and driving activity, rather than improving how people perform in role. In this sense, AI is often positioned implicitly or explicitly, as a form of displacement technology. Something that reduces reliance on people. But in practice, AI is proving to be far more effective as a tool for people.
Many solutions improve access, speed, or efficiency, providing information more quickly, automating simple tasks, or reducing administrative effort. These are valuable gains. But they do not necessarily change how people perform in the moments that matter. And without that change, organisations may see increased activity, but struggle to evidence meaningful impact on outcomes.
That’s where our approach is different. At Elephants Don’t Forget, we focus AI investment where it has the greatest impact – on people. We start by defining what success looks like in operational terms. Our data and performance experts work closely with customers to identify the outcomes that matter, and how they will be measured. From there, Clever Nelly is used to continually assess and build knowledge, confidence, and in-role competence, ensuring that learning translates into consistent performance.
Crucially, we stay close to that process, working alongside organisations to continuously measure, evidence, and refine the return being delivered over time.
Where to go next
The organisations seeing the strongest return from AI are not those deploying it most widely, but those applying it most deliberately. They are clear on the outcomes they want to achieve, focused on how performance changes in practice, and disciplined in how they measure and evidence impact.
We’ve brought together a set of real-world examples showing how organisations are using Clever Nelly to improve performance, reduce risk, and deliver measurable outcomes across their operations. From improving first contact resolution and reducing complaints, to accelerating speed to competence and increasing employee confidence, these examples demonstrate how AI investment can be directly linked to business impact.
Explore the e-book: 10 ways Clever Nelly improves performance