MCP Dev Summit North America

Enabling tens of thousands of engineers to work more efficiently with AI is a daunting task, especially when security is essential. At Amazon we’ve enabled our engineers to use hundreds of MCP servers for everything from ticket/issue systems to design/architecture guidance documentation. Along the way to achieving higher productivity, we’ve learned a few things about how to best rollout these kinds of massive changes. This session will share our essential learnings for MCP at enterprise scale. 

AI Agents at Uber may need to navigate a massive ecosystem of 1000s of services, handle sensitive data, and execute critical business logic. To enable this, we are moving towards an agentic future which leverages a unified Model Context Protocol (MCP) infrastructure to access real-time services.

We will share the architectural lessons learned from deploying MCP at an enterprise scale. We will dive into three key technical pillars of our strategy:

1. Protobuf-Driven MCP Servers: How we leverage existing services and protocol buffers to automatically generate MCP servers, providing safe and instant access to 1000s of microservices.
2. Derived Tools and Description Overrides: Why static tool definitions aren't enough for complex workflows. We’ll demonstrate how we allow developers to override and refine MCP tool descriptions via "derived tools," ensuring agents have the specific context needed for particular workflows.
3. Evaluate quality: How we evaluate quality of MCP tools, leveraged in our no-code Agent Builder, which is a tool that democratizes agent building at Uber.

MCP adoption in the enterprise is at an early stage and continues to grow rapidly. In the past year, Obot AI developed MCP gateways, registries, chat clients, and workflows for enterprise customers. In this talk, we will discuss the lessons we learned, the trends we observed, and the developments in MCP technologies we are particularly excited about for driving MCP adoption in the enterprise.

At Duolingo, we built an AI Slackbot using Claude's AI Agent SDK and the Model Context Protocol, connecting 180+ tools across 15+ MCP servers—GitHub, Jenkins, Sentry, Grafana, PagerDuty, Honeycomb, and more.

This talk covers:
1. Building and Standardizing MCP Servers: Patterns for stdio and HTTP transports with consistent configuration schemas.
2. Authentication and Role-Based Access Control: YAML-based tool allowlists gating sensitive tools by user role.
3. Evaluation Framework: A Python script runs parallel evaluation tests validating MCP tool usage, anti-hallucination checks, and response quality.
4. Real-World Challenges:
* Testing in production without synthetic environments
* Preventing customer PII from reaching AI models
* Performance optimization: Tradeoff between fast answer vs good answer
* AI hallucinating answers instead of calling MCP tools
* Wrong MCP parameters causing slow queries and failures
5. Live Examples: PagerDuty triage, Jenkins debugging, and incident investigation—all from Slack.

Connecting agents to tools is the easy part. Governing what they do with that access — what they can reach, under what conditions, with full visibility into what happened — that's the layer most enterprises are still building from scratch.

After 100+ enterprise MCP deployments across finance, HR, IT, and operations, one pattern keeps emerging: every organization is building the same missing layer. Not the model. Not the tools. The data plane that governs how autonomous systems operate inside the enterprise.

Workato CTO Adam Seligman will name the pattern and frame what that layer looks like at scale. Developer Advocate Zayne Turner will open the architecture behind a real deployment — showing how the data plane turns MCP from demo-ready integration into production-ready infrastructure.

Autonomy without control is risky. Data control without autonomy is stagnation. The enterprises that win will build both.

Building your first MCP agent is exciting. Building a platform where dozens of teams can safely deploy their own agents is a different beast entirely. This talk will share lessons learned scaling an MCP powered agent platform from a single prototype to a multi-tenant platform supporting diverse agent workloads.

Building your first MCP agent is exciting. Building a platform where dozens of teams can safely deploy their own agents is a different beast entirely. This talk will share lessons learned scaling an MCP powered agent platform from a single prototype to a multi-tenant platform supporting diverse agent workloads.