Agentic data science with the Data Agent Kit
cloud.google.com·7h ago
Data Science
Towards Data Science·Aug 26·5 min read
Is Agentic AI Just Automation?
TL;DR
Agentic AI often resembles traditional automation, functioning more like flowcharts than true autonomous agents. To create genuine agentic AI, developers must focus on building systems that exhibit real decision-making capabilities.
✦ Why It Matters
Evaluate your current AI projects to identify opportunities for integrating true decision-making capabilities.
Key Takeaways
- Focus on exception queues to find agentic opportunities.
- Differentiate between known unknowns and unknown unknowns.
- Leverage existing automation as tools for agentic systems.
- Start with low-risk domains to test agentic AI.
How It Works
Agentic AI systems operate by making real-time decisions based on data inputs that were not predefined. For example, in aviation maintenance, an agent can analyze fault history and operational costs to recommend actions that a human might not have considered.