I've been considering getting certified in UIPath, but now everyone is talking about "Agentic AI" and LLM-based workflows. Is traditional Robotic Process Automation becoming obsolete, or do I still need to learn the rule-based systems to be successful in an automation career?
3 answers
RPA isn't being replaced; it's being upgraded. Think of RPA as the "hands" that click buttons in legacy software and AI as the "brain" that makes decisions. In 2023 and 2024, we are seeing the rise of "Intelligent Automation." Companies still need RPA to interact with old ERP systems that don't have APIs. However, instead of a human writing 500 lines of rigid rules, we now use an LLM to interpret unstructured data (like a messy PDF) and then hand that clean data to the RPA bot. Learning both is actually the best way to future-proof your career and land senior roles.
Are you seeing more job openings for 'AI Automation Engineer' than for 'RPA Developer' lately? It feels like the titles are shifting even if the underlying work still involves a bit of both.
Traditional RPA is still vital for security and compliance. AI is too unpredictable for some financial tasks where you need a 100% deterministic result every single time.
Patrick makes a great point. For things like payroll or bank reconciliations, "creative" AI is a liability. You want the rigid, boring rules of traditional RPA for those.
Gregory, the shift is real. Most "RPA Developer" roles now expect you to know how to integrate OpenAI or Anthropic APIs into your workflows. If you only know how to scrape a screen and follow a fixed path, you'll likely find your salary ceiling is much lower than someone who can implement a "Human-in-the-loop" AI decision system. The industry is definitely pivoting toward the AI-first mindset.