Join our daily and weekly newspapers for exclusive content on the latest updates and industry-composure AI coverage. learn more
So far, many enterprises have started searching for AI agents and have started determining that deploying them is a viable option for their business. But many still match agents with some companies that have been for years: automation.
Leading to automation Uingpath Looks at agents and orchestrates the entire ecosystem – slightly different.
The company announced its new Uipath platform for agentic automation. However, it is clear that agents are not a new version of the robotic process automation (RPA); Rather, they are another device that can integrate with RPA to complete the enterprise workflow.
Umepath founder and CEO, Daniel Dines, in an interview told Venturebeat that agents could not be fully automated as they have been created today.
“The big problem with LLMS today is that they are nondeterministic, so you can’t run them directly into an autonomous fashion,” Dines said. “If you look at most of the implementation of agents, these are really chatbott. So we are going into chat, chat out of an agent that is in data, action out, where we orchestrates between agents, humans and robots.”
The key to the offering of the urepath is its AI orchestation layer, mestro. It oversees the flow of information from agents to an automation layer. Uipath described Maestro as a centralized observer, which “optimize automatic, models and complex business processes” and monitors performance.
Break up agents and automation
The Maestro user takes the signal and breaks the process in the managed stages to complete it. Instead of allowing agents to reach indiscriminate information, Dines said that Mestro has three stages.
- First, the agent takes the prompt, analyzes it, and advises how to complete the query.
- Next, a human user approves the recommendation.
- Then, an RPA equipment will perform the request and execute the recommendation.
Dines said that Mestro makes the workflow more transparent and accountable because a human remains in a loop and ends a rule-based RPA work. For Uipath, separating separate agents that take data to automatically take a recommendation that acts on the recommendation, ensuring that enterprises do not allow agents to reach their entire system.
“I think this is a very powerful way for enterprises to adopt agents. And look, in many discussions with customers, I think they are very well echoed because they are actually anxious about the unlimited agency of agents,” Dines said.
The Uipath also integrates the orcharging framework provider to offer multi-agent framework open with the provider Langchen. The platform for agentic automation also works with anthropic and microsoft framework, which is part of Uipath Google’s agent-to-agent protocol.
Not every agent is automation
Dines insisted that thinking about agents as a complete stack solution, where agents read data and then take action,
“Agents are transactions in nature as nondeterministic; they create effects on the underlying systems. No customer will not know that there will be a risk on it,” Dines said. “Transactions must be 100% reliable, and only automation can offer this type of reliability. So our solution is the best in those worlds.”
He said that “perhaps some future” agents AI “will become more reliable, and some actions that you can hand over to agents, but it should be a progress.”
Other people in the industry believe that agents are the next development of automation. In fact, the entire basis of agent AI is a system that does things on behalf of the user. A secondary goal for many is near the “environment” agents, where AI agents run in the background, work continuously for the user and consume people for any changes that require their attention.
However, Unepath still needs to make a case that its approach to agents is more effective than the all-in-one agent Prasad and cuts through the surrounding agents that do everything for users.
Companies such as Servicenow, Salesforce, Ruter and Microsoft have released all agentic platforms aimed at enterprise users. The author’s new platform depends on a self-developed model for autonomous agents.
Enthusiasts showed enthusiasm around the idea that AI agents can make their work very effective and automated many manual functions in companies.
