Moxie teamed up with OutSystems for a webinar tackling one of the biggest challenges engineering teams face today: moving beyond proof of concept to build AI agents ready for production. The full recording is now available.

On July 27, we joined OutSystems for "One App, Many Models & Engineering Mission-Critical AI Agents," a technical conversation on how to take agentic systems from prototype to enterprise environment without compromising governance, reliability, or control.

What was covered

The session opened with a real-world case: the modernization of an OutSystems application through a provider-agnostic AI layer, capable of routing requests across Azure, AWS, and locally hosted models. From there, the conversation moved into the engineering principles that make an AI agent truly ready for mission-critical use:

• How to build an AI architecture that doesn't depend on a single model provider.
• Guardrails and control mechanisms to keep agent behavior in check in production.
• Traceability as the foundation for auditability and trust.
• Continuous evaluation (evals) to monitor quality and performance over time.

Watch the full webinar

If you missed the live session, or want to revisit any of the technical points discussed, the full recording is now available to watch for free.

‍

‍