Your RPA team already has a head start on agentic AI (here's how to build on it)

There's a narrative forming in the market that agentic process automation (APA) requires a completely different talent profile than what most automation teams have today.

That narrative misses the point that people who have been building and running robotic process automation (RPA) programs for years already carry many of the instincts that agentic systems demand.

Process thinking, exception handling, understanding where decisions occur within a workflow, and knowing when something needs human judgment and when it can be automated. These are foundational to how agents operate, and they're skills that take years to develop through real-world practice.

You already have a headstart, but you might be wondering how to build on your foundation today.

Process fluency is the hardest skill to hire for
In this episode of the Agentic Edge Podcast, Mike Reynolds, a business and technology executive with over 20 years of experience in financial services at KeyBank, shared something that reinforced a pattern I see across the automation community.

When his team began building its automation capability years ago, they trained people from the line of business rather than hiring purely for technical skill. Roughly 80% of their builders came from business roles. They understood the processes deeply, could challenge assumptions, and earned trust with business partners because they lived the work firsthand.

That decision produced a team with something harder to replicate through hiring alone: cross-functional fluency. They could look upstream and downstream from systems, spot where processes actually broke down versus where documentation said they should work, and translate business pain into automation design.

If you think about what Agentic Process Automation (APA) requires, the overlap with traditional automation is significant. An AI agent reasons through a process, evaluates options at decision points, handles exceptions, and determines when to escalate. The people who already think in those terms have a meaningful advantage over someone learning process logic for the first time, regardless of how strong their technical skills are.

The skills that transfer and the gaps that remain
RPA builders bring strong process thinking, and that transfers well to APA. Mike's observation on the podcast was very direct: people who are really good at RPA also tend to understand agents, what they might do, and what they might not do.

The mental model of how processes flow, where they branch, and what happens when something unexpected occurs is directly applicable to designing agent behavior.

Where the gaps show up is in the measurement and evaluation layer. Traditional RPA has a relatively straightforward feedback loop. The bot completes the task, or it doesn't. Outputs are deterministic and testable.

Agentic systems introduce variability by design, which means teams need new skills around prompt evaluation, drift detection, confidence thresholds, and knowing when to trust an output versus when to escalate.

There's also a design thinking gap. RPA builders are accustomed to scripting specific steps in a defined sequence. Agentic solutions require thinking about goals, constraints, and reasoning paths rather than step-by-step instructions.

That shift from designing a fixed workflow to defining the boundaries within which an agent operates is a different kind of skill, and it's one that takes deliberate development.

The full spectrum is a design choice, not a migration
Let me be clear: the path forward is additive, not a replacement. RPA and APA occupy different parts of a spectrum, and strong automation teams will operate across the full range.

Some processes call for the repeatability and cost efficiency of a bot executing the same steps the same way every time. Others call for the reasoning and adaptability of an agent handling variable inputs and contextual decisions. And some call for a combination, with agents handling the cognitive layer and bots handling the deterministic execution through protocols like Model Context Protocol (MCP).

Mike framed this through a practical lens on the podcast: as a banker, when you think about the cost profile, RPA fits a narrow, efficient band, and knowing where each approach belongs is how you manage both cost and complexity.

The teams that understand that spectrum and can design solutions across it will deliver more value than those defaulting to one approach for everything.

Building on the head start
For automation developers and practitioners in the automation community, the message is straightforward. The skills you've built through robotic process automation (RPA) are genuinely valuable in the Agentic Process Automation (APA) world.

Process fluency, exception handling, understanding where things break down: these are your foundation, and they're harder to teach than any specific technical capability.

Building on that foundation means adding the layers that agentic systems specifically demand:

• Evaluation and observability skills so you can measure whether agents are performing as expected
• Design thinking around goals and constraints rather than fixed sequences
• Business fluency deep enough to connect what you're building to the outcomes that matter
• And the curiosity to keep exploring as the technology evolves

As Mike put it on the podcast, whatever you learn today, be willing to relearn in some form tomorrow.

Go Deeper
The full conversation with Mike Reynolds on the Agentic Edge Podcast covers how teams evolve from RPA builders to agentic practitioners, the leadership mindset required to invest in people alongside platforms, and why the organizations that build durable capability in their teams will be the ones to scale.