Getty Images/iStockphoto
Automation Anywhere adds reasoning to agentic AI platform
The robotic process automation vendor’s agentic process automation platform now has a reasoning engine that can understand enterprise context and help with agentic orchestration.
Robotic process automation vendor Automation Anywhere on Tuesday revealed that it has expanded its Agentic Process Automation (APA) system.
The RPA vendor introduced Agentic Process Automation (APA) a year ago and has marketed itself as a go-to vendor for agentic process automation. But it competes against others, including fellow longtime vendor UiPath, in the RPA market.
Agentic process automation uses AI agents and machine learning to automate business workflows in a way that is more active than traditional RPA, which is based on following preset rules. Instead, APA systems make decisions with AI agents and learn as they go.
At Imagine, Automation Anywhere’s user conference in Orlando, the vendor said the expanded APA system now includes a Process Reasoning Engine (PRE). The engine can understand enterprise context and work to achieve business goals using agentic orchestration. The APA system also includes Enterprise UI Agents, computer use AI agents that can operate a computer on their own, and Reasoning AI Agents that can plan, execute and learn once given a goal.
While the reasoning engine is now generally available, the Enterprise UI Agents are in preview.
Automation Anywhere's APA expansion comes after UiPath unveiled its Agentic Automation platform on April 30, and as the two established RPA vendors try to pivot to agentic AI. Blue Prism is another leading RPA vendor, and a slew of smaller RPA vendors are active.
Opportunity and threat
"Agentic AI is both a threat and an opportunity for RPA vendors," said Kashyap Kompella, CEO of RPA2AI Research. He added that it is an opportunity because it allows RPA vendors to evolve and become an orchestration platform for agents, mainly because they have the domain knowledge, trust of the enterprise because of the RPA vendors’ longtime presence and reliability, and established tooling infrastructure.
The threat comes because the agentic AI ecosystem is expanding quickly, with big tech vendors including Google, Oracle, SAP, Salesforce and others providing their own agentic orchestration products.
However, Automation Anywhere's addition of a reasoning engine is a step in the right direction for the vendor, Kompella said.
The PRE appears to combine enterprise process discovery and process analytics with newer AI models to optimize workflow, potentially enabling greater workflow automation than before, Kompella continued.
The RPA bots and agents work together, with RPA bots executing predefined steps and the agents using reasoning to interpret goals and perform actions or steps, he said.
"But note that enterprises everywhere are in the experimental phase when it comes to AI agents," Kompella said. "It's very early days."
Challenge of implementation
And the immaturity of the technology means putting it into action could be difficult, said Alan Pelz-Sharpe, founder of Deep Analysis.
"Implementing this is a challenge for anybody," Pelz-Sharpe said. "You need near-perfect understanding of the processes and the tasks, and perfect data, and very few organizations have that."
Moreover, he continued that RPA vendors and bigger platform providers are also fighting for the same spot as the preferred orchestration platform.
"That's where the battle will be who orchestrates all of these agents," he said. "If there's an army of agents out there, who's the orchestration layer? They will want to be that, because that's where the money is in the end."
On the other hand, RPA vendors like Automation Anywhere have a slight advantage because automation is their business, Pelz-Sharpe said.
"They've been studying these processes," he said. "They know about process mining. They know about task mining. They have partners who do this, so it's a natural evolution for them. Whereas the big platform vendors, well, they've got money and strength, but this is relatively new to them."
Esther Shittu is an Informa TechTarget news writer and podcast host covering AI software and systems.
