Decisions vs. Tasks: DHL's Prerit Mishra on What Agentic AI Actually Means in Operations

Agents that do tasks are automation with a chat interface. Agents that make decisions change how the business runs. DHL's Director of Data Quality gives leaders a clearer lens for where to invest first.

"Agentic AI" has become the most used — and most loosely used — phrase in enterprise technology. At AI Summit Seoul & Expo 2026, Prerit Mishra, Director of Data Quality at DHL, will offer a clarifying distinction that operations leaders can act on immediately.

The session

In "Decisions vs. Tasks: What Agentic AI Actually Means in Operations," Mishra separates two things that are routinely conflated. Agents that do tasks are, at heart, automation with a chat interface — valuable, but incremental. Agents that make decisions are something else entirely: they change how the business runs, and they demand different data, different governance, and different operating models.

The session offers leaders a clearer lens for where to invest first — grounded not in vendor demos but in the operational reality of one of the world's largest logistics networks.

Why it matters now

Agentic AI has overtaken generative AI as the boardroom priority, and budgets are following. But the failure pattern is already visible across the industry: companies deploying decision-making agents on task-grade data foundations and governance. As physical AI extends agents from software into warehouses, vehicles, and supply chains — a transition Korea's logistics and manufacturing giants are living right now — getting this distinction right is the difference between scaled impact and expensive pilots.

About the speaker

Prerit Mishra is a data and AI leader at DHL, where he builds the trusted data foundations that let AI deliver impact reliably across large, global operations. Previously Head of Data & AI for DHL Asia Pacific, he built and scaled the regional data and analytics capability and embedded AI into how the business runs. Across 15+ years in data, analytics, and AI spanning manufacturing and supply chain at Micron and DHL, he has helped move enterprise AI from pilots into operational reality. A frequent speaker across Asia Pacific, he holds a Master's in Automation from the National University of Singapore.

Key Takeaways

Two kinds of agents — why task agents and decision agents are fundamentally different investments, not points on the same roadmap.
The data and governance gap — what decision-making agents demand that most enterprise data foundations can't yet provide.
Where to invest first — a practitioner's framework for sequencing agentic AI in real operations.