AI has a value problem, not a technology problem

Organisations have spent the past two years investing heavily in artificial intelligence. New models have been deployed, pilot projects launched, and transformation programmes accelerated. Yet despite the investment, many leaders remain unconvinced that AI is delivering the value they expected.

The problem may not be the technology.

It may be the organisation's ability to recognise, communicate, and align around value once it has been created.

This question shapes the discussion in Episode 9 of The AI Values Podcast, where Edosa Odaro and Lindley Gooden examine what they describe as the “lost language of value.” Their discussion highlights a growing disconnect between those building AI systems, those funding them, and those expected to benefit from them. Everyone is talking about value, yet many appear to be speaking different languages.

Everyone sees the same thing differently

One of the most useful frameworks introduced in the discussion is the idea that organisations operate through three different languages.

The language of strategy.

The language of operations.

The language of influence.

Each serves a legitimate purpose. Senior leaders focus on growth, risk, and long-term direction. Operational teams concentrate on delivery, performance, and execution. Influential stakeholders often focus on adoption, alignment, and organisational priorities.

The problem is not that any of these perspectives are wrong.

The problem is that they often fail to connect.

"The language of strategy, the language of operations, but also the language of influence… it's almost like looking at an elephant from different perspectives, and everyone's in a slightly different picture."

As Odaro explains, people can be looking at the same initiative while reaching entirely different conclusions about its success. Technical teams may see sophisticated models delivering measurable improvements. Executives may struggle to identify business outcomes. Influential stakeholders may focus on entirely different priorities depending on their incentives and responsibilities.

The result is not necessarily failure.

The result is fragmentation.

Why AI value keeps disappearing

One of the most striking moments in the conversation concerns a statistic often cited in discussions about AI value.

Research suggests that only a small proportion of organisations can clearly quantify and communicate the value generated by their AI investments to boards and executive leadership teams. At the same time, many technical teams believe they are already delivering meaningful outcomes.

These two realities should not be able to coexist.

Yet they do.

The explanation offered in the episode is that value is often created but not translated. Technical teams describe predictive accuracy, model performance, and operational improvements. Boards are looking for evidence of changed behaviour, improved decisions, risk reduction, and measurable business outcomes.

Both groups are discussing value.

They are simply measuring it differently.

This distinction matters because organisational decisions are made through shared understanding. If different groups cannot agree on what value looks like, it becomes increasingly difficult to prioritise investments, allocate resources or assess success.

The challenge is not always proving value.

Sometimes it is making value visible.

The organisations that translate best will perform best

Throughout the discussion, Lindley Gooden repeatedly returns to the idea of translation.

Not translation between languages in the traditional sense.

Translation between organisational perspectives.

The argument is compelling because most organisations already contain specialists with deep expertise. Technical teams understand technology. Executives understand strategy. Operational leaders understand delivery. The challenge emerges when those groups attempt to communicate across boundaries.

Many organisations assume the solution is education. Some argue that executives should become more technically literate. Others believe technical teams should spend more time learning business strategy.

Odaro takes a different view.

"That value has got to be the thing that brings everyone together, and that language needs to be clear." 

Rather than expecting everyone to become experts in each other's disciplines, he argues that organisations need a clearer language of value that can bridge the gap between them. Value becomes the common reference point through which different groups can align without abandoning their own expertise.

This perspective has important implications for governance.

Organisations often invest heavily in technology capability while paying less attention to translation capability. Yet translation may be the mechanism through which value is ultimately recognised, funded and scaled.

Why more technical knowledge is not always the answer

One of the most interesting disagreements in the episode centres on leadership understanding.

A common argument within AI governance is that boards and executives need deeper technical knowledge. The logic is understandable. If leaders understand the technology better, they will presumably make better decisions about it.

Odaro challenges that assumption.

"Nobody wants your CEO to come downstairs to the technical department and micromanage everything you're doing." 

His argument is not that leaders should remain uninformed. Rather, it is that expecting executives to become technical specialists misunderstands the nature of leadership itself. CEOs, CFOs, and board directors already operate across a wide range of responsibilities. Their role is not to build models or manage technical architectures.

Equally, expecting data scientists and engineers to become business executives creates a similar problem.

The goal is not to turn everyone into the same type of professional.

The goal is to create a shared understanding of value.

This is a subtle but important distinction. Organisations often treat AI literacy as a knowledge problem when it may actually be an alignment problem. People can possess significant expertise and still struggle to work effectively together if they lack a common language for discussing outcomes.

The cost of misalignment is growing

The discussion ultimately points towards a larger organisational challenge.

Every year organisations invest more money, more time, and more effort into AI. Expectations continue to rise. Yet many of the statistics surrounding AI adoption point towards persistent dissatisfaction with results. Leaders feel value is not being realised quickly enough. Teams feel their work is not being recognised. Stakeholders become frustrated by the gap between promises and outcomes.

The risk is that organisations begin treating this as a technology problem.

In reality, it may be a communication and governance problem.

If value is being created but cannot be clearly articulated, trust begins to erode. If teams cannot explain outcomes in ways that resonate with decision-makers, support weakens. If leaders cannot connect investments to measurable organisational benefits, confidence declines.

None of these challenges is solved by better models alone.

They require organisations to become better at translating value across different perspectives.

The missing capability in AI governance

Episode 9 suggests that one of the most overlooked capabilities in AI governance is the ability to connect strategy, operations, and influence through a shared understanding of value.

This is not simply a communications challenge. It is a leadership challenge.

Organisations already possess technical expertise. They already possess strategic expertise. They already possess operational expertise. What is often missing is a mechanism that allows those perspectives to reinforce one another rather than compete.

The future of AI adoption may depend less on whether organisations can build increasingly sophisticated systems and more on whether they can create enough clarity for everyone involved to understand why those systems matter.

Without that clarity, value risks remaining hidden in plain sight.

Listen to Episode 9

Going deeper: why value has become a governance issue

One of the most important ideas in Episode 9 is that value is not an objective concept inside organisations. It is interpreted through different lenses depending on role, incentives, and responsibilities.

Technical teams often see value through performance improvements. Executives see value through outcomes and organisational impact. Stakeholders may see value through influence, adoption, or risk reduction.

The governance challenge is ensuring that these perspectives remain connected.

Without a shared language of value, organisations struggle to make informed decisions about investment, accountability and strategic priorities. The result is often frustration on all sides. Teams believe they are succeeding while leaders remain unconvinced.

This is why value should be treated as more than a measurement problem. Increasingly, it is becoming a governance capability. Organisations that can clearly articulate how value is created, measured and communicated are likely to make better decisions than those relying solely on technical performance metrics.