AI creates value when people trust it

AI discussions often begin with the technology. Organisations compare models, debate platforms and explore new capabilities. Yet the difference between successful AI programmes and unsuccessful ones rarely comes down to the technology itself. More often, it depends on whether people trust the data, the systems, and the decisions being made.

That is the challenge discussed in Episode 8 of The AI Values Podcast, where Edosa Odaro and Lindley Gooden speak with David Edem, Vice President Enterprise Data at AVEVA. Drawing on his experience leading data and AI transformation across complex industries, Edem explains why trust, stakeholder alignment and business outcomes remain the foundations of successful AI adoption.

The technology was never the hardest part

Many organisations assume that the primary challenge in AI adoption is technical implementation. The reality is often very different.

David Edem describes a large-scale engineering information challenge involving the acquisition of energy assets containing decades of operational history. The organisation inherited vast quantities of engineering drawings, project documentation, seismic data, and operational records spread across multiple systems. Consolidating this information into a trusted and usable form would traditionally have required years of effort and significant investment.

AI fundamentally changed the economics of the project.

By applying automated document analysis and data extraction techniques, the organisation was able to dramatically reduce the time and resources required to create a consolidated source of truth. A programme that had previously appeared to require several years and millions of pounds became achievable within a single year and at a fraction of the cost.

What makes the example particularly instructive is that AI was not working alone. Engineers, document controllers, and domain experts remained deeply involved throughout the process. Their role was not to compete with the technology but to validate, review, and improve its outputs.

The lesson is important because it challenges the common assumption that AI replaces expertise. In practice, many successful implementations depend on combining machine efficiency with human judgement. The technology accelerates the work, but trust is established through people.

Why business outcomes matter more than AI

One of the strongest themes in the conversation is the importance of framing AI initiatives around outcomes rather than technology.

Edem explains that when he enters executive discussions, he rarely begins by talking about AI. Instead, he focuses on the business problem that needs solving and the value the organisation is trying to create. The technology becomes relevant because it contributes to the outcome, not because it is innovative.

This distinction may sound simple, but it has significant implications for governance.

Boards and executive teams are rarely interested in AI for its own sake. Their priorities are growth, resilience, operational performance, risk reduction, and long-term value creation. AI only becomes strategically meaningful when it advances those objectives.

The conversation highlights a common failure pattern within organisations. Technology teams become enthusiastic about capabilities while business leaders remain unconvinced about relevance. The result is often a disconnect between innovation activity and organisational priorities.

Successful AI programmes close that gap by linking technology directly to business outcomes. They answer practical questions about value, risk, and performance rather than focusing exclusively on technical features.

Trust cannot be automated

Throughout the episode, trust emerges as a recurring theme.

Organisations need to trust their data. Leaders need to trust the outputs being presented to them. Employees need confidence that the systems they are using are producing reliable information.

Generative AI complicates this challenge because it introduces a new type of risk.

"AI wants you to be happy. It gives you good results and it makes stuff up."

The observation captures one of the defining governance challenges of the current AI era. Large language models often generate highly persuasive responses regardless of whether those responses are correct. Confidence and accuracy can appear indistinguishable, particularly for users who lack the expertise to verify outputs independently.

This is why Edem consistently emphasises the role of human oversight. AI may perform the majority of the work, but people remain responsible for quality assurance, validation and accountability. In his experience, AI often delivers 80 to 95 per cent of the outcome, while human expertise ensures that the final result is trustworthy.

For leaders, the implication is clear. Responsible AI is not simply about deploying better systems. It is about designing processes that ensure confidence is supported by evidence rather than assumption.

The most important people are often the undecided ones

When organisations discuss AI adoption, attention tends to focus on supporters and sceptics. The conversation rarely considers a third group that often has significant influence over outcomes.

The undecided.

Edem describes these individuals as neither enthusiastic advocates nor active opponents. Instead, they delay decisions, request additional information, and hesitate before committing to change. While they may attract less attention than vocal supporters or critics, their influence can be substantial.

This observation highlights an often-overlooked aspect of AI governance. Adoption is not simply a technical process. It is a human process.

Before entering a boardroom or presenting a major initiative, Edem argues that leaders should identify potential supporters early, involve them in the conversation, and bring them on the journey. Building trust before formal decisions are made often proves more effective than attempting to win support during a high-stakes presentation.

The principle applies well beyond AI. Organisational change succeeds when people understand why it is happening, how it will affect them and what value it is expected to create.

Why storytelling is becoming a leadership skill

One of the most memorable observations from the discussion concerns communication.

"We that do technology and AI need to be better at telling stories."

Behind the statement sits an important leadership challenge. Technical expertise alone is rarely sufficient to drive adoption. Organisations are made up of people with different priorities, levels of understanding, and concerns about change.

For many leaders, the ability to translate technical possibilities into business language is becoming increasingly important. Stakeholders need to understand not only what a technology does but also why it matters, how risks will be managed, and what outcomes can realistically be expected.

Storytelling, in this context, is not about marketing.

It is about alignment.

The organisations most likely to succeed with AI are often those that can create a shared understanding of purpose across technical teams, operational functions and executive leadership.

AI has lowered the barrier to entry

The discussion also addresses one of the most significant changes brought about by generative AI: accessibility.

Historically, advanced AI work was concentrated among highly specialised teams with deep technical expertise. Today, access to powerful AI capabilities has expanded dramatically. Professionals from non-technical backgrounds can experiment, learn, and create value using tools that would have been inaccessible only a few years ago.

Edem shares the example of a senior AI leader whose career journey began in customer success rather than engineering. Within a relatively short period, self-directed learning and practical experimentation helped that individual move into a global AI leadership role.

The example illustrates how rapidly the landscape is changing.

Lower barriers to entry create new opportunities for innovation and learning. They also increase the importance of governance, critical thinking, and digital literacy. As AI becomes more widely accessible, organisations will need stronger foundations for ensuring that these tools are used responsibly and effectively.

The future of AI depends on human judgement

The conversation repeatedly returns to a simple but important truth. AI can automate tasks, accelerate analysis and create efficiencies that would previously have been impossible. It can help organisations unlock value from information at a scale that was once prohibitively expensive.

Yet technological capability alone does not determine success.

The quality of data, the strength of oversight and the ability of leaders to build trust all play a critical role in shaping outcomes. As organisations move beyond experimentation and into large-scale deployment, these factors become increasingly important because they determine whether AI-generated insights can be translated into meaningful decisions.

Technology creates possibilities. Organisations create value when people trust those possibilities enough to act on them.

Listen to Episode 8

Going deeper: trust is becoming a strategic capability

One of the most valuable insights from Episode 8 is that trust should be viewed as a strategic capability rather than an abstract concept.

AI systems do not create value simply because they exist. They create value when organisations are confident enough to integrate them into decision-making, operational processes and strategic planning. That confidence depends on far more than technical performance.

It depends on data quality, transparency, accountability, and communication. It depends on whether leaders can explain how systems work, where risks exist, and what safeguards are in place. Most importantly, it depends on whether people believe that the outputs generated by AI can be trusted in the context of real-world decisions.

For boards and executives, this creates a leadership challenge. Building trust is no longer solely a cultural responsibility. Increasingly, it is becoming a prerequisite for effective AI governance and long-term value creation.