Presentation: Architecting AI Systems for the Messy Reality of Enterprises: Why Agentic Compute is the Missing Layer
Deutsche Telekom's LMOS shows what enterprise agents look like at scale—and why most platforms are building it wrong.
Why it matters
A practitioner presentation on production agentic architecture reveals the missing layer between chatbots and operational intelligence: ephemeral agents, platform abstractions, and Agent Definition Language (ADL) as the bridge between organizational silos and autonomous workflows.
The key facts
11 to knowDeutsche Telekom deployed LMOS as production agentic platform
Focus on ephemeral agents for operational intelligence (not persistent chatbots)
Agent Definition Language (ADL) as core abstraction for multi-team coordination
Tool sprawl consolidation into platform abstractions identified as key blocker
Organizational fault lines (between teams/domains) as architectural constraint
Case study: Deutsche Telekom's LMOS (agentic platform in production)
Key pattern: ephemeral agents replacing persistent tool chains
Architecture concept: Agent Definition Language (ADL) as abstraction layer
Problem addressed: organizational fault lines in enterprise agent deployment
Transition from chatbots to operational intelligence systems
Concept of 'agentic compute' as missing infrastructure layer
Go to the source
InfoQ AI/MLinfoq.com
Publisher excerpt: Arun Joseph shares real-world insights on scaling enterprise agentic platforms like Deutsche Telekom’s LMOS. He discusses bridging organizational fault lines, replacing tool sprawl with core platform abstractions, and moving beyond basic chatbots to operational intelligence systems through…