Low Code Approach Episode 98: Traditional Automation vs AI-Powered Agents (w/ Carsten Groth)
Aug 5, 2025
Carsten Groth, a Global Black Belt in the Power Platform at Microsoft, discusses the evolving landscape of automation. He uncovers why AI-powered solutions are outpacing traditional robotic process automation, emphasizing the game-changing role of large language models. The conversation also highlights innovative strategies for integrating AI into existing systems, transforming processes like supplier onboarding. Carsten encourages businesses to rethink their automation approaches and embrace new technologies for enhanced efficiency.
AI Snips
Chapters
Transcript
Episode notes
LLMs Simulate Human UI Workflows
- Computer use automation uses LLMs to simulate human UI interactions like filling forms and attaching files.
- It represents a new class of automation beyond traditional RPA by applying natural language and AI to UI tasks.
RPA Was A Bridging Technology
- RPA acted as a bridging technology for systems without APIs by simulating mouse and keyboard actions.
- AI-driven approaches are less deterministic and can handle natural language inputs more flexibly than classic RPA.
Use Agents To Extend RPA
- Combine agents with RPA where needed by letting agents call RPA for legacy UI tasks.
- Use natural language in agents to describe new fields so the model adapts without repeated developer intervention.