AI research lab Anthropic recently said the Claude 3.5 Sonnet large language model (LLM) can now interact with computers. It can view a screen, move a cursor, click buttons, and type text, allowing it to perform tasks like filling out forms, navigating websites, and interacting with a wide range of software programmes.
It is similar to what those Robotic Process Automation (RPA) companies promised a decade ago.
This agentic approach from Anthropic makes many of us wonder about the potential implications for RPA companies like Automation Anywhere, UiPath, and tech giants like Microsoft and IBM, which have RPA solutions for enterprises.
Today, the industry is on the cusp of AI agents, which promise similar automation capabilities. So, this begs the question: What does the future hold for RPA?
The Future of RPA
In a previous interaction with AIM, Mark Geene, senior VP and general manager of product at UiPath, said that RPA bots primarily handle tasks from high repeatability to moderate complexity. More straightforward tasks are increasingly managed by autopilots or task-level coding.
However, as we approach the peak of complexity, tasks become less repeatable. They will require an AI agentic approach to handle more nuanced and variable situations.
“By categorising tasks into these quadrants based on complexity and repeatability, we can better select the appropriate tools for each job. This approach allows us to leverage automation where feasible and deploy agents where human judgement and adaptability are essential,” Geene said.
Param Kahlon, EVP & GM of automation and integration at Salesforce, shared the same sentiment. At its flagship event, Dreamforce, the CRM company announced Agentforce, a new platform that lets enterprises build AI agents within minutes.
At the sidelines of Dreamforce held in San Francisco, Kahlon told AIM that autonomous agents also do not mean the end of RPA technology.
“RPA agents were designed to automate repetitive, tedious tasks, such as transferring data between systems when APIs aren’t involved. In contrast, autonomous agents process information more like humans, adapting to situations and making decisions based on changing conditions, enhancing efficiency and effectiveness in workflows,” he said.
This was the general consensus. However, the new Anthropic update changes the dynamics completely. Now, AI can be used to automate repetitive, tedious tasks—things that RPA technology is being deployed to do. Companies such as Asana, Canva, Replit, and the Browser Company have already begun exploring its potential.
Even though Anthropic has claimed that the ‘Computer use’ feature is in the beta phase, it will most likely get better. So, does this make RPA technology obsolete? Not really, at least for now.
RPA Companies are Jumping in the AI Agent Bandwagon
Companies such as UiPath and Automation have thousands of customers worldwide. UiPath, alone, has over 10,000 customers who have deployed RPA technology at an enterprise level.
While they would definitely look into the new Anthropic features, it would be unlikely that they would disrupt their existing systems with something which is still being tested.
Many of these customers would definitely test these newer systems; however, enterprises often choose the ones that bring more value, not the new cool technology.
Moreover, these companies making RPA bots have also started leveraging AI Agents. UiPath co-founder Daniel Dines believes that in the future, AI will do around 80% of what humans do at work.
At its FORWARD user conference in Las Vegas, UiPath also announced the integration of Anthropic’s Claude 3.5 Sonnet into its products. In fact, Dines went on to say that AI agent-led automation is ‘Act Two’ for UiPath.
( UiPath co-founder and executive chairman Daniel Dines)
This statement signifies that RPA companies are jumping on the AI Agent bandwagon and will soon offer AI-agent-led automation for enterprises.
Automation Anywhere, a competitor to UiPath, has also not shied away from AI agents. Earlier this year, it introduced a new feature that allows customers to create custom AI agents that can learn from enterprise data, make informed decisions, and act responsibly within the enterprise system.
Automation at Scale is not Easy
RPA technology has existed for a while. UiPath, founded in 2005, has been helping enterprises automate workflows for almost 20 years. Yet, there remains room to automate more. This is because automation is not easy.
“The biggest lesson that we had throughout these years is that automation is really hard. It’s very hard to deploy automation at scale,” Dines himself said while addressing the keynote session at FORWARD 2024.
This is because enterprise processes are often complex making enterprise-level automation challenging. Moreover, many enterprises still leverage legacy systems adding to complexity.
Yet, RPA has managed to run thousands of automations in a reliable way over the years. Could an AI agentic approach solve these problems? It remains to be seen.
Dines also said that RPA bots are deterministic, whereas AI agents are not. “It’s extremely difficult to use it ( AI Agents) in the context of an enterprise workflow because enterprise workflows need to be reliable, deterministic,” he said.
In the upcoming months, enterprises will evaluate the effectiveness of AI agents and their potential benefits. It remains to be seen whether enterprises can derive real value from these agents. Hence, RPA bots will continue to support businesses for the time being, although their future remains uncertain.
The post Did Anthropic Just Kill RPA Bots with Computer Use? appeared first on AIM.
