AI Show and Tell - New York

Agents are easy to demo. Running them is the hard part.

Global AI New York is back at Microsoft Lafayette for an evening about what happens after the demo: evaluating agents, governing them, and running a team of them in production.

Three talks, real systems, real failure modes. Plus a light dinner and time to talk to the people building this stuff.

Agenda

18:00 Welcome and light dinner

19:00 Your New Teammates: Architecting a Multi-Agent Workforce
Henk Boelman
Inside hamba.nl, a live Dutch travel magazine run by seven Foundry-hosted agents that cover the full editorial pipeline from draft to published article. Scoping, orchestration, MCP, handoffs, and where the human stays in control.

19:30 From Findings to Fixes: How SCOPE and Azure Cosmos DB Improved AI-Generated Apps
Jay Gordon and Cedric Vidal
Eight AI-generated applications, six stacks, three months of testing. Broken SDK usage, expensive queries, unsafe data patterns, and what it took to actually fix them.

20:00 From Intent to Enforcement: Open-Source Evaluation and Governance for AI Agents
Mehrnoosh Sameki
ASSERT turns plain-language expectations about agent behaviour into executable evaluations. ACS carries them into production as portable policies. Together they close the loop from intent to enforcement.

Who should come
Developers, architects and anyone putting agents in front of real users.

Free to attend, seats are limited.

Full details and speaker bios: https://globalai.community/e/783bfa20

Hosted by Global AI New York, part of the Global AI Community.