AI Strategy vs AI Reality

Stephen Redmond on Leadership, Literacy & Shadow AI

The Gartner hype cycle recently placed generative AI in the trough of disillusionment. After the initial rush of excitement, organisations are confronting a sobering reality: many AI pilots fail to deliver promised results. Stephen Redmond, former Head of Data and Analytics at BearingPoint and Data and AI Innovation Lead at Accenture, has seen this pattern before.

In this candid conversation with Maryrose Lyons, Redmond cuts through the hype to examine what separates successful AI implementations from costly failures—and why leadership literacy matters more than ever.

The Hype Cycle Reality

Redmond acknowledges that AI is experiencing predictable backlash. The initial rush to adopt generative AI saw organisations launching pilots without clear objectives. Now, reports suggest up to 95% of these initiatives fail to move beyond experimentation.

But Redmond argues this is not evidence that AI lacks value. Rather, it reveals a fundamental gap between AI strategy and AI reality. Successful implementation requires more than technology—it demands organisational readiness, clear use cases, and sustained leadership commitment.

The Shadow AI Problem

One of Redmond's most pressing concerns is shadow AI—unsanctioned use of AI tools by employees without IT oversight. As generative AI becomes more accessible, staff increasingly use consumer tools for work tasks, creating security, compliance, and data governance risks.

Redmond argues that organisations must address shadow AI not through prohibition but through education and provision of approved alternatives. When employees understand AI capabilities and limitations—and have access to vetted tools—they make better decisions.

Leadership Literacy: The Critical Gap

Redmond identifies AI literacy among leadership as the single biggest predictor of implementation success. Leaders who understand AI's capabilities and constraints can set realistic expectations, allocate appropriate resources, and recognise meaningful progress.

Conversely, leadership that treats AI as magic—or expects immediate transformation without investment in data infrastructure and change management—sets initiatives up for failure.

From Pilot to Production

The path from successful pilot to production deployment requires addressing several challenges:

Data Infrastructure: AI systems require clean, accessible, well-governed data. Many organisations discover their data estate is not AI-ready only after launching pilots.

Change Management: AI transforms workflows and job roles. Without clear communication and training, adoption stalls.

Measurement: Organisations must define success metrics before launching initiatives. Too many AI projects chase novelty rather than measurable business outcomes.

Implications for Professional Services

For engineering, architecture, and construction firms across Ireland and the UK, Redmond's insights carry particular weight. These industries face pressure to adopt AI for competitive advantage, but must do so within strict regulatory and safety frameworks.

Professional services firms need AI governance frameworks that enable innovation while managing risk. This requires leadership that understands both the technology and the specific compliance requirements of their sector.

Conclusion

AI's move into the trough of disillusionment is not a signal to abandon investment—it is an invitation to invest more thoughtfully. Success requires leadership literacy, clear strategy, and disciplined execution. For organisations in Ireland, Dublin, Athlone, and across the UK, the question is not whether to adopt AI, but how to do so responsibly and effectively.

Want the full conversation? Watch the Chatting GPT episode on YouTube here: https://www.youtube.com/watch?v=cdL-f2WiBmc