The unavoidable connection between AI readiness and AI ROI

Without these answers, AI investment can become driven by what’s technically possible rather than what’s strategically valuable. And this approach is typical at the start of most AI programmes. Everyone’s focused on adopting new models, introducing automation, and experimenting with agents, rather than connecting actions to business goals. Perhaps, rightfully so. Many times we don’t know what we don’t know until we’ve experimented a little. But then we need a plan to move from experimentation to value realisation.

It’s important to remember that a technical capability doesn’t automatically lead to a business outcome. Automating a workflow might appear to reduce the time that’s spent on a task, but if your people have to verify AI outputs and correct errors — much like one would verify a junior team member’s work — that gain might not be as significant as you think; it could even equate to a productivity loss.

A baseline gives you a reference point against which you can set realistic expectations, prioritise investment, and measure whether AI is improving the outcomes that matter to the business.

Many of the items above are questions engineering would ask for any software developed in a normal product lifecycle. Creating value with AI is no different; we should still ask the same questions. Using AI does not bypass these business-critical questions. These costs don’t necessarily appear in your investment business case, but they can determine whether an AI initiative creates a genuine productivity gain or simply moves work somewhere else.

Our research into ‘Understanding the human cost of AI transformation’ highlights this issue. Many organisations assume AI automatically saves time, but our research suggests otherwise. It revealed that:

AI can redistribute effort from execution to verification, creating a hidden operational and psychological burden – what we call the “verification tax” – when adoption isn’t designed around how people actually work. And that matters for ROI.

Rather than putting the brakes on AI adoption until every capability is perfect, you need to understand where readiness matters most and address the gaps that could undermine the value you’re trying to create.

AI technology becomes a real enabler when it’s supported by the right training, processes, and organisational culture. That means teaching your people how AI fits into their work, where they can trust it, when they need to check it, and how they remain accountable for the outcomes.

Research from MIT's NANDA initiative has reported that 95% of organisations are getting zero return from their generative AI investments, despite substantial spending and widespread experimentation.

The reason is that they’re rushing to deploy AI while many employees are still dealing with the human challenges created by previous waves of digital transformation. Our research shows that although AI was intended to simplify work, in some cases it’s causing stress, information overload, context switching, and a growing sense of disconnection from the purpose of work.

If AI is introduced as part of a deliberate change to workflows, roles, skills and ways of working, it has a much better chance of removing friction and creating meaningful value. That’s why we believe better AI ROI starts with a clearer understanding of current maturity, an honest assessment of the constraints that could affect value, and a plan to intentionally close the gaps that matter most — and leveraging a battle-tested AI organisational change management approach that works synergistically with a cohesive AI Strategy Programme, much like what we do with any other major digital transformation programme.