Daniel Dines, co-founder and CEO of UiPath.
Robots have evolved. From early black and white movies (Metropolis) almost 100 years ago, robots finally became a reality and started helping to weld rivets in automotive engineering factories around the 1970s. Then, robots evolved into robotic humanoids, and a Japanese company invented a proof-of-concept walking robot, but these are still hardly part of our homes. Then, throughout the post-millennium era, as the era of robotic process automation began to flourish, software robots, or “bots,” emerged, giving software systems new autonomous capabilities to perform not only maintenance functions but also useful and productive business tasks.
Despite the lip service to robotics' brief history, the current landscape of robotics and RPA may still be a mix of software and hardware, but first and foremost, it is a balanced equation of software.
Closing the AI Gap
John Kelleher, UiPath's regional vice president for the UK and Ireland, is keen to push the idea of robotics beyond many of the existing notions of automation.
“We all know about the impact of artificial intelligence. The problem is there's a gap between the promise of AI and the reality of how it's deployed today,” Kelleher said. “McKinsey's analysis on this area shows that companies are thinking too narrowly about the business case for how AI should be applied in the workplace. It's not just a matter of technology choice, there's also a significant education aspect to overcome, coupled with a data architecture that allows change management to have a real impact on how IT and business functions work together. Aiming to close the gap between the promise of AI and the reality of how it's deployed, we're looking at how companies can build the operating models of the future and really create new ways of working.”
Today, AI might be considered pervasive, meaning it can potentially be applied to every aspect of a business. Kelleher says that in this sense, AI shouldn't be viewed as a single application deployment point. Instead, it should be seen as a fabric that can be applied to specifically defined solutions that span the end-to-end operating model across the modern enterprise.
It’s clear that UiPath’s work in the UK and Ireland region mirrors and resonates with its work in North America and the rest of the world. UiPath CEO and founder Daniel Dines has more to say about robotics than most. Speaking of the “dark ages of AI and RPA” (referring to as recently as 2016), Dines says the team realised early on that the two fields were symbiotic.
What is Agent-Based Technology?
“The future of AI and RPA will be agentic, so let's take a step back and consider what the term means and why it applies to the next age of automation. Simply defined, the term agent is the ability of an AI system to control and manage the flow of a business process,” explained Dines, speaking to reporters and analysts in London this month. “Many human actions – thinking, moving, blinking, etc. – are done automatically, without us expending any cognitive capacity. I mention this because human intelligence is being driven by these automated routines.”
To provide more background, we ourselves can be defined as agents. Humans in call centers or business people at any level are called agents when they perform a function. These days, it is common to talk about software agents, i.e. software (or complete computer programs) that perform prescribed tasks for human users, machines, or other virtually defined entities that form part of a workflow system. So, in this context, a human-agent team is a collaborative environment made up of interacting humans working in tandem with an AI system.
This clarification is necessary because Dines is painting a picture of a world in which human agents increasingly hand over tasks to AI and RPA, which raises the question: what aspects of our lives will we continue to be responsible for, rather than what can be left to computers?
Left-Right Brain Thinking
Dines asks us to think more about automation. If you're thirsty in the middle of the night, he says, you just grab a bottle of water from the fridge and drink it. This action is all automatic, learned, and done without thought, so (as we said before) this is all left-brain automatic thinking. Left-brain thinking is about the robotics where humans create and maintain automation, and it's where structured, logical, efficiency-oriented, systematic processing happens.
The right hemisphere is responsible for creative intuitive thinking that applies adaptability and deals with ambiguity. It is also where all large scale models of behavior are autonomously decided, with the ability to manage adaptive behavior.
To bring this element of intelligence into RPA, you need AI models that can understand exceptions as they occur and better learn and react to real-world ambiguities. This is the point where you can add “agent skills” to RPA and start to bring the intelligence from what UiPath calls self-healing robots into dynamic planning and dynamic learning in your business.
The Future of Our Agents
“In the future of agents, we see AI evolving to a level where it can handle around 80% of what humans do at work. This is when AI starts to take on more spontaneous (perhaps unstructured, unexpected) roles in business (or applications like healthcare, or any industry),” said UiPath's Dines. “One might ask, what would it take to offload 100% to an AI agent? I don't think AI is at that point right now. However, there are some instances or tasks where offloading is possible (e.g., self-driving cars), but only in deployment scenarios with relatively controlled, defined, and safe environments. So, while there is work to be done, the future is definitely very exciting.”
The company's comments come at a time when the company has made some straightforward updates to its platform. UiPath is currently introducing several new features to its platform designed to embed generative AI more deeply into the UiPath Business Automation Platform. UiPath Autopilot, for developers and testers, uses generative AI and natural language processing in UiPath Studio to create workflows, generate expressions, and help build automations.
“New features include text-to-workflow conversion: developers simply describe their automation idea in natural language and Autopilot creates the initial workflow; text-to-formula conversion: With Autopilot, developers no longer need to remember the exact syntax and structure of a formula. Developers can write what they need in natural language and let the AI generate the correct formula; text-to-code conversion: Generates code from natural language descriptions to reduce the deployment time of automation projects,” the company said in a tech product statement.
The preview of the UiPath plugin and integration with Copilot for Microsoft 365 provides an integration that allows users to automate end-to-end business processes with their colleagues directly within Microsoft Teams. Customers can access a library of pre-built automations to complete common repetitive tasks as well as specialized automations for function or industry-specific tasks. Users can also discover and run automations developed by their own companies.
Workers and agents united
Robotics is on the rise, expanding the scope of automation, evolving adaptability, evolving scope of functionality, and expanding its scope of application in business. This proliferation is driving more and more work to be taken over by AI-powered software agents. This is not a bad thing, but rather a good thing. As one UiPath spokesperson said, “If you can apply RPA and AI to your role and find a way to fully automate your work, you're clearly a smart person, and that's the surest way to get promoted.”
But whatever the outcome of AI agents and RPA automation applied to business, the changes here are sure to make management consultants happy: If we can automate that (sometimes precarious) role as well, then AI will have truly come of age.
