From Assistant to Agent: Rethinking AI and the Role of Service Design
When I was studying service design, I imagined myself probably working on customer journeys, touch-points, and occasional workshops where sticky notes covered the wall, when I graduate. What I didn’t expect was to land up in a start up building around an agentic AI, a technology that doesn’t just follow instructions but actively reshapes the workflows and relationships we design around. It was while listening to IBM’s AI in Action podcast about how they transformed HR with agentic AI that a powerful analogy landed with me about the difference between a footballer’s assistant and an agent.
The player’s assistant is responsible for helping with logistics like booking flights, carrying kit bags, arranging dinners, etc. They make life smoother but are fundamentally limited. An agent, on the other hand, negotiates contracts, brokers partnerships, and steers the player’s career with strategic foresight. They don’t just execute, they influence outcomes. This distinction and autonomy is similar to what separates traditional AI from agentic AI. Old systems acted like assistants, waiting for prompts and handling tasks on demand. Agentic AI has the power to remember, reason, and take initiative across contexts with the ability to act on behalf of the user and shape results in ways that extend beyond the immediate task.
For a service designer, that shift builds on the very foundation of how we work. Instead of designing for a passive tool that users command, we are now designing for a collaborator that contributes, adapts, and sometimes even leads. And this reframing naturally leads to a bigger question of “who is the user really?”.
IBM’s adoption of agentic AI through their “Client Zero” philosophy helped me see this clearly. Before offering their AI externally, they tested it internally in HR, positioning their employees as their first audience. That approach not only ironed out bugs, it also challenged the narrow idea that the “end customer” is always the only primary user.
But the lesson goes further than that. If you zoom out, you realise that users aren’t just individuals at all anymore. They exist within an ecosystem that could cover employees , managers, partners, customers, even regulators and stakeholders who all interact with or are impacted by the system. What excites me most is that this perspective broadens the canvas of design.
We are no longer mapping human-to-human static journeys, we are orchestrating services where agents behave more like teammates than tools, and where “users” can be anyone touched by that interaction. It forces us to think less in terms of tasks and more in terms of value creation.
It’s not only about whether AI helps people do their jobs more efficiently, but also about whether it builds trust, reduces friction, empowers collaboration, and creates clarity across.
As a fresh service design graduate, it feels both daunting and exhilirating to step into this new terrain. But it is also a moment of opportunity to redefine what service design means in the age of agentic AI , and to help organisations see that creating value is not about solving for one user, but about shaping a system where everyone, from employees to customers, benefits from collaboration. That is how we move from being AI users to becoming true AI value creators.
