When offices around the world emptied in early 2020, few anticipated that the main challenge wouldn’t be remote work itself, but rather an explosive growth in manual operations. “What used to be resolved through a quick conversation by the coffee machine now turns into dozens of emails and filled-out forms,” says Yuri Astafiev, an AI and business process automation expert, serial technology entrepreneur whose projects have gained recognition in the international professional community.
This challenge has been particularly acute in sales departments. “Think of a sales manager who used to collect leads at conferences and business meetings. Now they have to spend hours combing through LinkedIn and other platforms, manually building databases of potential clients,” explains Yuri, who has over eight years of experience in building development teams and has implemented more than 40 systems utilizing artificial intelligence.
The scale of the problem is confirmed by McKinsey’s global survey conducted in July 2020: two-thirds of executives reported increasing investments in automation and artificial intelligence. “But the key question is why aren’t existing automation solutions handling this challenge? The answer lies in how we fundamentally think about automation,” notes Astafiev.
“Traditional RPA solutions were designed for a different era,” Yuri explains. “They excel at automating stable, repetitive processes in large companies. A typical robot can log into a CRM, export a report, distribute it to a list – and do this 24/7. But as soon as we step outside strictly defined scenarios, the system fails.”
Indeed, leading RPA developers position their products as “digital workforce.” However, this has proved insufficient in the new environment. “The issue is that modern business processes are becoming increasingly fluid and dynamic. Interfaces change, new fields and forms appear, companies constantly optimize their processes. And each such change requires robot reconfiguration,” says the expert.
“We are on the threshold of a new generation of RPA systems,” Astafiev believes. “The essence of the change is that we’re moving from simple action repetition to intelligent interface interaction. It’s like the difference between a robot that can only click buttons at predetermined coordinates and a robot that understands the purpose of these buttons.”
Machine learning and computer vision synergy plays a crucial role in this transformation. “Modern algorithms allow the system to perceive interfaces holistically, understand relationships between elements and their functional purpose. This radically changes the approach to automation,” explains the expert.
“What’s particularly important is that such systems can operate without traditional programming. It’s like teaching a robot to think like a human. It’s not just executing a predetermined sequence of actions, but understanding what it’s doing and why,” Astafiev elaborates.
According to Yuri, the next stage of RPA evolution is inextricably linked with artificial intelligence. “We need systems that don’t just follow predefined algorithms but understand the context of their actions. Imagine that instead of rigidly programming each step, you can simply show the system what a ‘Submit’ button or ‘Email’ field looks like, and it will learn to find these elements even if their appearance or location changes.”
Specifically, applying computer vision and deep learning methods enables the system to recognize interface elements regardless of their exact location or appearance. “It’s like teaching a robot to ‘see’ and ‘understand’ rather than just following rigid screen coordinates,” explains Yuri. This approach has already proven effective in practice – while working on Robotic Tape, Astafiev’s team successfully applied computer vision and machine learning technologies to create a system capable of adaptively interacting with web interfaces.
“But recognition alone isn’t enough,” continues the expert. “It’s critical to ensure system reliability and stability in a production environment. RPA robots often handle confidential data and critical business processes. Any failure can have serious consequences.”
The expert believes that the key to ensuring reliability lies in a multi-level action validation system. “Each robot action must go through several levels of verification. The system must not only ‘understand’ what it’s doing but also be ‘aware’ of the potential consequences of its actions.”
“When we talk about scalability, we face another interesting challenge,” notes Astafiev. “The system needs to work equally efficiently with ten or a thousand parallel processes. Meanwhile, it’s important to optimize resource usage and ensure rapid recovery after failures.”
“When discussing the future of RPA systems, it’s important to understand that we’re on the verge of a qualitative leap,” reflects Yuri. “Automation is no longer just a tool for performing routine tasks. It’s becoming an intelligent assistant capable of making decisions in complex situations.”
According to the expert, in the coming years, we’ll see the emergence of RPA systems capable of handling significantly more complex scenarios. “Imagine a system that doesn’t just collect data about potential clients, but analyzes their digital footprint, evaluates partnership potential, and even generates personalized proposals. And all this without human intervention.”
Particularly interesting prospects are emerging at the intersection of RPA and creative tasks. “We’re already seeing how artificial intelligence helps create content, generate images, write code. The next step is integrating these capabilities into corporate processes through RPA systems,” explains Astafiev.
“But the most important change will not be about technology, but about the approach to implementation,” emphasizes the expert. “We’re moving from a model that required serious technical expertise to configure each process, to systems that can be configured with just a few clicks. This will make the technology accessible to a much wider range of companies.”
“Ultimately, the future of RPA isn’t just about technological evolution, but a fundamental change in how we think about automation,” Yuri concludes. “Artificial intelligence allows us to move from mechanical repetition of actions to creating truly intelligent systems capable of learning and adapting.”
“But it’s important to remember that technology is always a tool, not an end in itself,” Astafiev concludes. “The key is understanding real business needs and being able to offer a solution that truly simplifies people’s work. In this sense, artificial intelligence isn’t a magic wand, but a powerful tool that we need to learn to use properly.”
In an environment where digital transformation is becoming essential for business survival, the development of intelligent RPA systems appears to be the inevitable next step in the evolution of corporate automation. And those companies that can master these technologies first will gain a significant advantage in the post-pandemic world.
Featured image credit: Steve Johnson/Unsplash
