Tech Disruptors AWS and How it is Building Enterprise Agents
40 snips
Nov 10, 2025 
Swami Sivasubramanian, VP of Agentec AI at AWS with over 20 years of experience, joins Anurag Rana to delve into the future of enterprise AI agents. They discuss the shift from simple chatbots to agents capable of autonomous planning and task completion. Swami highlights the importance of security, data access, and governance in deploying these agents. He also explores challenges in moving from prototype to production, the significance of context and memory for personalization, and how agents can modernize legacy systems, paving the way for greater efficiency.
AI Snips
Chapters
Transcript
Episode notes
AgentCore Shortened Production Time
- AWS built AgentCore to provide secure, scalable foundations so teams can deploy agents faster.
- Swami gave examples like Ericsson and Cohere Health using AgentCore and Transform saving hundreds of thousands of hours.
Solve Production Gaps Before Deploying
- When moving from prototype to production, address containerization, multi-tenancy, identity, and observability up front.
- Use runtime isolation, identity integration, tool sandboxes, and memory to avoid brittle deployments.
Agents Can Query Multiple Silos Securely
- Agents can dynamically retrieve relevant data across clouds and on-prem systems using connectors plus identity-aware access.
- They perform agentic retrieval-augmented generation to combine multiple data sources without heavy custom pipelines.