From RPA to Agentic AI

Agentic AI takes your tools from just doing to understanding and organizing with AI.

From executive bot to self-thinking AI assistant

Fifteen years ago, Pegamento began as a pioneer in Robotic Process Automation (RPA), at a time when the idea that software “robots” could take over routine tasks from humans still sounded revolutionary. We built bots that performed repetitive tasks: retyping data between systems, filling out forms, processing documents. Automation then meant: mimicking human mouse clicks and keyboard strokes. Useful, efficient, but limited.

As the years progressed, so did ambition. Artificial intelligence (AI) found its way into our solutions and changed the nature of what a “process” actually is. We were no longer just automating how something was to be done, but also when, why and for whom.

The big advantage? Pegamento has 15 years of hands-on experience with real customer cases – at a time when we didn’t even name AI to customers, because it simply wasn’t a relevant label.

The Pegamento Process Assistant was one of the first tangible examples of this shift: not a bot that merely follows instructions, but an assistant that thinks, anticipates and adapts to the context. Think of a colleague who understands what needs to be done and why – and converts that into action independently.

And yet that was only the beginning.

References for Agentic AI

Municipality of Amsterdam

RPA speeds service delivery, reduces steps and increases citizen satisfaction.

Achmea

Smart desktop automation speeds up customer contact and significantly reduces IT costs.

From instruction to initiative

Today we are in the midst of a new phase: that of agentic AI. Where RPA once began by mimicking human behavior, and was later enriched with AI to understand and predict, we are now talking about agents that can act independently. No more “executive hands” dependent on a workflow, but digital assistants with purpose, memory and initiative.

An agent knows: “I have to answer this customer question”, looks up the right information independently, decides which steps are necessary, executes them and learns from the outcome. Like a colleague who doesn’t have to tell you what to do every second. No more scripts, but small digital entities with a real role, a sense of context and the ability to make choices – within clear boundaries (guardrails). Policies and ethical guidelines are essential in this regard.

This shift is fundamental. It changes how we work, how we distribute work, how we design systems and how people and technology work together. At Pegamento, we don’t see this as science fiction, but as a logical next step. We know the landscape of processes – from control systems to AI models – and know what it takes to make that landscape passable for agents. Not as an isolated innovation, but as a continuation of a journey we have been building for 15 years.

Who are the forerunners in Agentic AI?

The frontrunners in agentic AI are organizations where processes are intensive, error-prone and time-consuming – and where the stretch is on. After all, without urgency, no change.

Public service providers, such as municipalities, healthcare institutions, implementing agencies and health insurers, are experiencing increasing pressure: changing legislation, high workloads and citizens who expect quick, clear answers. Here, an AI agent can independently review files, prepare decisions or handle contact moments – without each step requiring human intervention.

Also in the financial sector adoption is accelerating. Banks, leasing companies and insurers spend a lot of time on compliance, customer identification, reporting and process verification. There is no more room for waiting or manual work there. Agentic AI handles complete tasks here: from analyzing documents to identifying risks and making preliminary decisions. Did you know that, on average, one customer interaction in this sector leads to seven follow-up tasks?

And then there are the commercial organizations – from e-commerce to telecom – where scalable, personalized customer experience is crucial. Here, an agentic AI interprets a customer query, retrieves data, executes a transaction and provides feedback within seconds. 24/7, with no wait time.

Ready for the next step

The typical users? Organizations with many repetitive processes, complex rules and a high dependence on human capacity. Not to replace people, but to give them space back. So they can focus on the real work: contact, empathy, strategy and growth.