When machines join the team

By Michael Bradshaw, Global Applications, Data and AI Practice Leader at Kyndryl

For decades, enterprises have trusted managed technology service providers to keep their mission-critical operations running efficiently and securely. The people behind this work have served as guardians of infrastructure, keeping financial markets open, power grids stable, commercial planes flying, and hospitals functioning without pause. The job may not be glamorous, but it is indispensable.

The infrastructure underlying these critical components of the economy is complex and requires nuanced thinking and accountability. In these environments, human experience and judgment is essential.

With advances in artificial intelligence, these roles have been evolving, and now, with agentic AI, they will undergo even more profound transformation. AI is shifting from being a background tool to becoming something more consequential: a digital colleague.

In this new blended workforce, AI agents don’t simply automate repetitive tasks, they reason, coordinate, and adapt in real time. They troubleshoot, learn from past outcomes, and collaborate with other systems, taking on many tasks once thought to require a human touch. For enterprises accustomed to long-established linear models of service delivery, this represents a seismic shift.

But it would be a profound mistake to assume AI alone can shoulder the weight of managing the world’s most critical tech infrastructure. These are environments where failure is not an option, trade-offs can be agonizing, and accountability always rests with human leaders. An airline may deploy AI agents to anticipate weather-related disruptions, but when thousands of passengers are stranded across multiple hubs, it takes human judgment — informed by context, empathy and experience — to decide how to respond.

And therein exists a core reality: AI agents bring speed, scale and adaptability. Humans bring judgment, context and trust. In the future workforce, bringing all those qualities to bear will be what drives managed services forward.

But in order for this new paradigm to work, the integration of AI into existing systems must be more than thoughtful. It will have to be intentional, governed and human-centered. That’s because it involves a radical reshaping of the technology stack, skills and decision-making models that allow humans and machines to truly thrive together. It’s a cultural challenge, and it requires organizations to foster a new mindset where humans and machines learn from each other and adapt together. It means developing new communication norms, new governance models and new ways of measuring success.

Ultimately, the promise of agentic AI is not that it will replace human workers, but that it will amplify them — extending their reach, enhancing their judgment and enabling them to manage complexity at scale. Agentic approaches are designed to augment human teams, reduce cognitive burden and accelerate decision-making.

Agentic’s advantages

The benefits of harnessing AI agents are multifold. They can monitor the network and systems in real time and predict potential disruptions. They can upgrade security by detecting anomalies and threats, instantly triggering automated responses. And they can analyze and report on energy and data use and allocations.

As they continue to progress, AI agents can spot and correct inefficiencies while making recommendations for improved performance. And with human oversight and engagement, the agents can leverage organizational information to identify silos and areas of duplication.

Human managers guide all of this, providing critical guidance and supervision and stepping in when tough decisions are necessary. For example, AI agents are used in hospital systems to predict surges in demand and allocate scarce beds to patients with the most acute health needs. But doctors and nurses still supervise, signing off on agentic decisions and overriding them when necessary, applying human judgment and sensitivity when making difficult choices.

Setting the right tone

Not every enterprise is ready to take advantage of agentic AI, and leaders need to prepare their organizations for what is both a technological and cultural change. As enterprise leaders evolve their AI use from pilot programs and one-off experiments into core functions and decision-making processes, they must prepare their workforces for what many will perceive to be a jarring and even threatening shift. Managers can ease the transition by leading through example, demonstrating the advantages of working with agents and sharing the results. They can also engender trust by establishing firm guardrails for the use of AI and emphasizing the role of humans in controlling it.

Of course, not every employee will be ready, and enterprises will need to invest in teaching and training workers to take advantage of this new paradigm. The success of these efforts will rest on the trust employees have in their employer and the benefits they can perceive from their new AI colleagues.

Technology is advancing at breakneck speed, and it’s opening new opportunities and improving the quality of life for billions around the world. But with this next era comes new, complex systems and an ever-growing dependence on them. Undergirding it all is digital infrastructure, the mission-critical backbones that can never fail. Humans and machines will share responsibility for managing that infrastructure.

Michael Bradshaw

Global Applications, Data and AI Practice Leader