Enforcing Real-Time Guardrails for Interoperable AI Agents on the Financial Desktop
AI and vibe-coding make building powerful desktop tools astonishingly fast - but how do you govern what an AI agent is actually allowed to do across your applications?
Generative AI and vibe-coding have unlocked unprecedented development speed. In financial services, these new AI tools can query sensitive databases, discover capabilities dynamically, interact with legacy front-ends, and execute trades or workflows on behalf of users. Interoperability via standards like FDC3 makes these AI applications exponentially more powerful - but it also creates significant governance, security, and compliance challenges.
Join Chris Watson and Rob Moffat for an immersive, hands-on workshop to explore the intersection of desktop interoperability and AI safety. Using FDC3-enabled platforms—including the open-source FINOS FDC3 Sail - you will build vibe-coded tools, integrate AI agents, transform policy rules into enforceable real-time guardrails, and then actively test those constraints to see if they hold under pressure.
Workshop Description
This highly practical session is built around a simple principle: move fast with AI without losing control. Participants will spend the afternoon building, connecting, and attempting to stress-test governed AI agent workflows in a safe, browser-based desktop environment.
- Facilitators: Chris Watson, Alpha FMC & Rob Moffat, FINOS
WHO SHOULD ATTEND?
This workshop is designed for technologists building, managing, or securing next-generation financial desktop environments:
- AI & Agentic Workflow Engineers: Who need to map LLMs and autonomous agents to structured, deterministic FDC3 desktop intents.
- Front-End & Desktop Developers: Looking to integrate AI agents safely into micro-frontends, containers, and interoperable workflows.
- Desktop Architects & Security Leads: Responsible for setting policy guardrails, context-sharing boundaries, and identity controls across financial applications.
- Product Managers & Business Technologists: Eager to leverage AI to eliminate manual re-keying and streamline trader and analyst workflows without risking non-compliance.
What You Will Walk Away With:
- Hands-On FDC3 Sail Deployment: Practical experience configuring FDC3 Sail—the open-source, web-based training desktop agent—to orchestrate inter-app communication.
- Deterministic AI Agent Tooling: Strategies for converting FDC3 intents and contexts into structured API tools that LLMs and AI agents can invoke reliably.
- Enforceable Policy Guardrails: How to place runtime compliance constraints directly in the execution path of AI actions to prevent unauthorized data exposure or operations.
- Adversarial Testing Experience: A real-world perspective from attempting to "break" governance controls, revealing where AI agents fail and how to patch vulnerabilities.
WORKSHOP PREREQUISITES & REQUIREMENTS:
- Laptop Required: You will be running browser-based tools, writing/editing code, and executing agent tests.
- Technical Baseline: Familiarity with modern web development (JavaScript/TypeScript, HTML/CSS), basic API concepts, and general familiarity with LLM/AI prompt or tool-calling mechanics.
Logistics & Application
- Date: Tuesday, November 3, 2026
- Time: 9:00 AM – 12:00 PM EST
- Location: New York, NY (Venue details provided upon registration)
- Cost: Free for registered attendees (Standard Value: $1,000)
Note: Seating is strictly limited to ensure a collaborative, hands-on environment. Priority will be given to individuals from FINOS Member financial institutions and regulated industries.
APPLY FOR SEAT HERE
(if the form does not show up below because of company blocking, please try this link)