The future of RPA ties to AI agents

Robotic process automation is accelerating toward a crossroads, with IT leaders and experts increasingly debating the technology’s future. Some IT leaders believe that more powerful and autonomous AI agents will eventually replace the two-decade-old AI precursor technology, while others predict that AI agents and RPA will work hand in hand to achieve new levels of automation.

RPA, which is still widely used in several industries, including manufacturing, healthcare, and finance, will likely be reconsidered in the years ahead as organizations begin to deploy agents and other next-generation AI tools. But the technology’s days may not be numbered, if vendors and IT leaders can create new ways of automating business processes by linking up RPA and AI agents.

Moreover, some IT leaders see RPA living on, having carved out a vital enterprise niche.

“RPA is still relevant for automating rule-based, repetitive, and redundant tasks, especially in industries where there is a big downside for an error like banking, insurance, and healthcare,” says Arjun Bali, staff data scientist at Rocket Mortgage. “It is not yet being replaced, but augmented with AI, allowing for smarter decisions within workflows.”

Adaptable vs. cost

While AI agents offer a powerful, adaptable, and autonomous approach to automation, good, old RPA, with its predictable outcomes, still has a place, says Shae Khan, AI research scientist at the IBM MIT AI Lab.

AI tools may eventually replace some RPA deployments, but RPA can be cheaper and faster to deploy, while being less prone to errors than most AI tools, Khan adds.

Still, many organizations will soon use AI to augment, and in some cases, replace traditional RPA, he notes. AI agents will be used to handle complex and dynamic tasks that require decision-making capabilities, while RPA will continue to be used for repetitive, rule-based processes.

“Unlike RPA bots, that follow predefined rules, AI agents are learning from data, making decisions, and adapting to changing business logic,” Khan says. “AI agents are being used for more flexible tasks such as customer interactions, fraud detection, and predictive analytics.”

Kahn sees RPA’s role shifting in the next three to five years, as AI agents become more prevalent. Many organizations will embrace hyperautomation, which uses multiple technologies, including RPA and AI, to automate business processes.

“Use cases for RPA most likely will be integrated into broader AI-powered workflows instead of functioning as standalone solutions,” he says.

RPA growth still happening

RPA vendors such as UiPath that have pivoted to add AI agents to their portfolios still see a future for RPA, even as they tout the advantages of the newer technology. UiPath, in a report released in January, contends that agents will connect to RPA processes because of the efficiency and reliability of the older technology. IDC projects that RPA spending will more than double between 2024 and 2028 to reach $8.2 billion, the report notes.

Most of UiPath’s customers are still heavily reliant on RPA, says Chris Radich, public sector CTO at the automation provider. While there’s a budding narrative that AI will replace RPA, those making the prediction don’t understand the fundamental differences in the two technologies, he adds.

“RPA isn’t dying — it’s evolving,” he says. “We’ve tested various AI solutions for process automation, but when you need something to work the same way every single time —without exceptions, without interpretations — RPA remains unmatched.”

Radich and other automation experts see AI agents eventually controlling RPA bots, with various robotic processes in a toolbox for agents to choose from.

“Today, we build separate RPA workflows for different scenarios,” Radich says. “Tomorrow, with our agentic capabilities, an agent will evaluate an incoming request and determine whether it needs RPA for data processing, API calls for system integration, or human handoff for complex decisions.”

Organizations will need an orchestration approach to solve the major problem of automation sprawl, he adds.

“We have dozens of RPA bots, various API integrations, and manual processes that handle exceptions,” he says. “An orchestrating AI agent could manage this complexity by understanding the full landscape of available capabilities and routing work appropriately.”

Foundation layer

For the moment, the top use cases for RPA are in structured back-office processes, such as processing invoices, reconciling data, and generating reports, says Lei Gao, CTO at conversational AI platform vendor SleekFlow.

“Those [processes] aren’t going away overnight, especially in legacy-heavy sectors,” he adds. “But we’ll see a steady decline in new RPA implementations for those tasks, as more businesses adopt AI-first workflows.”

Gao foresees a change in how RPA will be used in the next three to five years. “RPA will likely be the foundation layer — still valuable, but increasingly invisible, embedded within AI-powered orchestration systems,” he claims.

CIOs and other IT leaders need to pay attention to these automation changes because the growing relationship between AI and RPA will affect maintenance, adaptability, and control over the automation infrastructure, Gao says.

“The ability to adapt in turn creates new possibilities — but also trade-offs,” he says. “AI agents add new levels of abstraction, which may make behavior more difficult to follow or to debug, in particular in regulated domains where the ability to explain the decisions made is important.”

A shift toward a more AI-powered automation environment will require CIOs to have deep domain knowledge and sound strategic judgment, Gao says.

“The higher-level question is, ‘What degree of autonomy and adaptation do we desire in our automation layer — and what transparency are we willing to lose in order to gain velocity and scale?’” he says. “The shift isn’t simply moving from one set of tools to another. It’s rethinking the philosophy of automation for the age of AI.”