In finance, making every step agentic can create unnecessary risk. Calculations, reconciliations, materiality thresholds, and validation rules are often better handled deterministically. AI becomes useful where judgment and investigation are needed: connecting evidence across files, exploring unusual movements, and explaining what changed.
In this code-along webinar, Jayeeta Putatunda, Forward Deployed AI Engineering Lead at Turing, will build a Financial Controller Agent with OpenAI Codex that reviews monthly financial results, identifies material variances, investigates exceptions, and prepares findings for human review. The session will show how to combine deterministic controls and agentic workflows in one system, with evaluation gates and human escalation built in. The goal is not to add AI everywhere, but to use it selectively where it creates real value without weakening the controls that financial processes depend on.
Key Takeaways
- Learn how to use OpenAI Codex to power AI agents that work with real-world financial data.
- Understand where deterministic rules are more appropriate and where agentic reasoning adds value.
- Build a financial controller agent from scratch that reads reviews monthly financial results,
