Everyday AI Podcast – An AI and ChatGPT Podcast From Automation to Agents: Why Weak Data Makes AI Guess
28 snips
Dec 11, 2025 
In this chat, Ed Makosky, Chief Product and Technology Officer at Boomi, dives into the intriguing world of agentification in AI. He explains how AI agents differ from traditional automations, emphasizing their ability to adapt to faulty inputs. Ed discusses the risks of weak data leading to poor outcomes and outlines strategies for improving data quality and governance. He also reveals how agents can revolutionize enterprise workflows, using expense reports as a practical example of their efficiency and potential. Tune in for insights on balancing automation and data integrity!
AI Snips
Chapters
Transcript
Episode notes
Agents Will Guess When Data Is Weak
- Agentified workflows will produce outputs even when data or configuration is wrong, which can lead to guesses or lies.
- This makes weak data more dangerous with agents than with brittle traditional automations.
Prioritize Data Quality Before Agentification
- Build agentic automations on a strong foundation of clean, governed data to avoid harmful actions.
- Invest in data quality first, because bad data in leads to worse agentic outputs.
Expense Reports As An Agent Example
- Ed uses expense reports to illustrate moving from rigid workflows to adaptive agentic processes.
- Agents can simplify submission, learn approver habits, and only escalate truly out-of-policy cases.