Chatham Financial, a leader in capital markets advisory, has integrated OpenAI models like Codex and GPT-5.6 to transform its workflows. This shift offers valuable insights for beginners learning to build software with AI. By focusing on outcomes, Chatham demonstrates how AI can enhance efficiency while maintaining accuracy and auditability.
Reengineering Trade Validation with Codex
Chatham’s Process Zero initiative reengineers workflows to maximize efficiency. A key example is trade validation, where AI tools reduce review time from 30 minutes to under 4 minutes. Using Codex, Chatham built an application that gathers transaction evidence, compares key terms, and flags discrepancies for review.
The application undergoes rigorous validation against experienced reviewers to ensure accuracy before expanding automation. This approach highlights the importance of balancing automation with human oversight, a critical lesson for beginners.
Tip
Always validate AI-driven tools against human performance before scaling automation.
Scaling Employee-Built Applications
Chatham Vibes, an internal platform, empowers employees to build applications tailored to their work. These applications leverage GPT-5.6 Terra by default, with GPT-5.6 Sol as an optional upgrade. This flexibility allows teams to choose the most suitable model based on their needs.
Examples include tools for reviewing maturing-cap trades, preparing pricing workbooks, and producing fixed-income rate sheets. These applications demonstrate how AI can support diverse workflows, providing inspiration for beginners exploring AI-driven software development.
Building Chatham Onyx with Codex and OpenAI Models
Chatham Onyx, the firm’s next-generation capital markets operating system, integrates assets, debt, and derivatives into a governed environment. Codex aids across the development lifecycle, from planning to testing, accelerating the path from concept to product capability.
The platform leverages multiple OpenAI models:
| Model | Use Case |
|---|---|
| GPT-5.6 Sol | Complex tasks requiring high accuracy |
| GPT-5.6 Terra | Default for AI features |
| GPT-5.4 | Cost-effective for simple analysis |
| GPT-4.1 | Non-production testing |
This tiered approach ensures cost-efficiency while maintaining performance, a strategy beginners can adopt when managing resources.
Enhancing Expert Capacity with AI
AI tools at Chatham perform structured comparisons, organize evidence, and identify exceptions, freeing experts to focus on high-value tasks. This shift underscores the role of AI in augmenting human expertise rather than replacing it.
For beginners, this highlights the importance of designing tools that complement human skills, enabling users to apply their judgment where it matters most.
Frequently Asked Questions
How does Chatham ensure the accuracy of AI-driven tools?
Chatham validates AI tools against experienced reviewers and real transactions before scaling automation, ensuring reliability.
What models does Chatham use in its AI applications?
Chatham uses GPT-5.6 Sol, GPT-5.6 Terra, GPT-5.4, and GPT-4.1, selecting models based on task complexity and cost-efficiency.
Lessons for Beginners Building AI-Driven Tools
Chatham’s approach offers several key takeaways for beginners:
- Focus on Outcomes: Design workflows around desired results, identifying where human judgment is essential.
- Validate Rigorously: Test AI tools against human performance before scaling automation.
- Leverage Flexibility: Use multiple models to balance cost and performance.
- Augment Expertise: Build tools that free experts to focus on high-value tasks.
By adopting these principles, beginners can create AI-driven tools that enhance efficiency while maintaining reliability and accuracy.
Read next
- What Faster Video Diffusion on TPUs Means for Beginners Building AI
- How to Choose and Use GPT-6 Models for Building Software with AI
Want to try all of this hands-on? Start with the free Claude Code from Zero course.
Source
Based on OpenAI’s announcement, “Chatham scales its capital markets expertise with OpenAI”. Written for people learning to build with these tools.