For many, AI still feels like just another tool in the digital transformation arsenal. But Yuma sees it differently, and so does its Managing Partner, Michel Herquet.

In this Trailblazer episode, Michel challenges the conventional view of transformation, arguing that AI marks a fundamental shift: from systems people must adapt to, toward systems that adapt to people. With multi-agent architectures, collaborative AI, and a more human-centric approach, Yuma is helping organisations embrace a future where AI is not a tool but a thinking teammate.

Michel, could you start by explaining what you see as the core difference between traditional ‘digital transformation’ and the new wave of ‘AI-driven transformation’?

It’s tempting to treat AI as just another tool in digital transformation, like cloud or ERP systems. But AI is fundamentally different. Technologies like Large Language Models aren’t just built, they’re discovered. Even the experts who train them don’t fully understand their inner workings, which introduces uncertainty.

This stems from machine learning’s adaptive nature. As we pursue general intelligence, explainability diminishes, and system design must become more iterative and experimental, unlike the structured methods of traditional digital transformation.

Yet that’s a strength. AI learns and evolves, enabling more human-centred, bottom-up transformation tailored to people, data, and context. It demands a mindset shift, but for those who embrace it, the rewards include smarter systems, deeper insights, and greater adaptability. This isn’t just a continuation of digital transformation. It’s a new, more human chapter.

You emphasise the concept of the ‘AI agent.’ How would you define an AI agent, and why is it so crucial to understanding this next phase of transformation? 

At its core, an AI agent combines reasoning capabilities, communication (like natural language processing), and the ability to act (such as updating a database or triggering a workflow). More than that, agents are built for collaboration, working together to accomplish tasks no single agent can achieve alone.

This collaborative model makes AI agents central to the next phase of transformation. And yes, humans are agents too, which is key to building hybrid teams.

In classical digital transformations, people often adapt to standardised processes and tools. In an AI-driven approach, machines adapt to humans. Can you expand on how this shift changes the relationship between technology and people within organisations?

For decades, IT systems enforced standardised processes to optimise efficiency, requiring people to adapt to rigid tools. But today, agility and resilience matter just as much as efficiency. AI offers an alternative: systems that adapt to people and changing conditions.

This shift redefines our relationship with technology, from using tools to collaborating with intelligent systems. It allows organisations to build more flexible, context-aware solutions that evolve alongside human needs.

Many firms outside specialised industries still struggle to demonstrate tangible value from AI. What, in your opinion, has changed in recent years that might finally help companies overcome these challenges?

The big shift isn’t just in performance. It’s in the economic model. Previously, AI projects began with costly, time-consuming data collection, creating high barriers to entry. Now, foundational models like GPT offer pre-trained intelligence that can be fine-tuned or adapted quickly.

This “model-centric” economy means companies no longer need vast custom datasets to get started. These general-purpose models bring AI within reach for a broader range of use cases, industries, and budgets, making value creation far more accessible.

You talk about forming ‘hybrid’ teams composed of both human experts and AI agents. Could you walk us through a real or hypothetical example that illustrates how these ‘multi-agent’ setups can boost performance and innovation?

This is already a reality. In contract generation, AI agents now assist humans by collecting data from third-party sources, APIs, and internal systems, then helping draft accurate documents. In call centres, AI supports customer service by automating background tasks like pulling up data or logging cases, letting human agents focus on conversations.

In both cases, the human-AI collaboration enables faster, more personalised, and more consistent results, unlocking new levels of productivity and innovation.

When CIOs look to integrate AI into their organisations, what strategic steps or principles would you recommend they follow, especially if they’re just starting to explore AI-driven changes?

We recommend a four-step approach to implementing AI effectively across your organisation. It begins with inspiration, where the focus is on understanding AI’s potential and defining a shared strategic vision that aligns all stakeholders. The next phase is design, in which AI initiatives are identified and prioritised, with a continuous evaluation based on return on investment, always asking: What does it cost? What will it deliver? Following this is the build phase, where resources are committed with clear expectations about outcomes and the potential to scale. Finally, in the run phase, the emphasis shifts to adoption. Tailored change management, hands-on support, and continuous monitoring are essential to embed AI successfully and sustainably.

One key factor often overlooked is talent. Insourcing is ideal but slow; outsourcing risks fragmentation. That’s why we advocate for a Team-as-a-Service (TaaS) model, agile, outcome-driven teams embedded within your organisation for speed, flexibility, and effective knowledge transfer.

The more AI is integrated, the more important governance, resilience, and ethical considerations become. How do you advise CIOs to set up frameworks that maintain transparency and trust while still encouraging innovation?

Hybrid human-AI teams and multi-agent systems offer built-in mechanisms for governance and resilience. For example, data access can be managed at the agent level, and critical tasks can include human-in-the-loop checks. If one agent fails, others can step in, boosting fault tolerance.

Because humans remain involved in decision-making, they’re empowered to challenge AI outputs. This dynamic fosters ethical oversight, transparency, and trust, without sacrificing speed or innovation.

There’s a growing focus on social and environmental impact. How can organisations ensure that their AI initiatives align with these broader responsibilities without sacrificing profitability or competitive advantage?

The current automation-first mindset, aimed at replacing humans, has led to limited gains and growing resistance. It’s also environmentally costly due to oversized models used indiscriminately.

We advocate a different paradigm: designing hybrid Human-AI systems where humans aren’t just safeguards. They’re essential contributors to performance. This approach not only aligns with social and environmental responsibility but also drives deeper value creation and long-term success.

Shifting to an AI-centred model typically involves new skill sets and a different mindset. What are the key capabilities organisations must develop, and how might they go about nurturing these skills within their teams?

Understanding how to interact with AI is important, but often overemphasised. Skills like prompt engineering are helpful but not transformational.

What matters more is developing collaborative and critical thinking, communication skills, and the ability to manage distributed systems. These aren’t new skills, but they need to be realigned with AI strategies.

Most organisations already offer leadership or creativity training. The key is integrating these efforts with AI adoption so teams can truly thrive in this new context.

Finally, if you could share one overarching message with technology leaders about the potential of AI as the next industrial revolution, what would you want them to walk away remembering above all else?

Don’t think of AI as a polished product. It’s not finished—it’s a raw material we’ve only begun to shape. Like steam or electricity in past revolutions, it takes vision, courage, and creativity to realise its full potential.

The organisations that embrace this complexity and build boldly will shape the future.

As Michel reminds us, AI is no longer something you deploy. It's something you collaborate on.

This new phase of transformation is not about replacing people, but about designing systems where people and AI agents elevate each other. This means rethinking strategy, governance, and even the role of technology itself.

 --