Chatham Financial builds capital markets tools with OpenAI Codex – Unite.AI

OpenAI published a customer case study on October 2, 2026, describing how Chatham Financial, a capital markets advisory firm, uses OpenAI’s Codex and GPT-5.6 models in its advisory workflows, including a trade validation application that the firm says reduced review times from approximately 30 minutes to less than four.

The case study lists Chatham as a small to medium-sized business in the financial sector in North America, with ChatGPT, Codex and the OpenAI API as products in use. The firm works with clients on complex capital markets decisions. According to the case study, Chatham uses Codex to build internal and customer-facing tools and GPT-5.6 to power AI capabilities, work that spans employee-built applications, its next-generation capital markets operating system, Chatham Onyx, and its reengineering consultancy service, Process Zero.

CEO Matt Henry said the company starts with the desired outcome, identifies where judgment matters and then plans how to deliver that outcome with currently available capabilities.

Commercial validation under process zero

Through Process Zero, Chatham redesigns workflows based on results. For each workflow, the company establishes the minimum inputs and tests needed, establishes where human judgment remains essential, and specifies how AI and tools built on AI should support the work.

Commercial validation is a prime example. Chatham’s Data Integrity and Controls team protects the accuracy of transaction data by confirming that every record in its systems matches what the customer authorized and what was actually executed. Using Codex, Chatham has developed a trade validation application that collects evidence to support each transaction, checks key terms and flags discrepancies for review, and the company is comparing the application’s results with those of expert auditors.

“In initial measurement, the application reduced review times from approximately 30 minutes to less than 4 minutes. Equally important, we are validating its performance against real transactions and expert reviewers before expanding automation. Ultimately, we will be able to produce easily auditable and more accurate results faster.” —Alex Nordlinger, co-head of Chatham’s AI consulting practice

According to the case study, Chatham plans to extend the application to additional business types and further automate the workflow while maintaining adequate controls and professional oversight.

Applications created by employees on Chatham Vibes

Chatham’s customer-facing AI work grew out of its own operational experience, the case study says. Employees apply ChatGPT and Codex in research, analysis, writing, software development and other daily tasks. Employees have also become builders through Chatham Vibes, an internal platform on which they create applications suited to their work. By default, AI features in Vibes applications run on GPT-5.6 Terra, and individual apps can be configured to upgrade to GPT-5.6 Sol.

Such applications increasingly support customer-facing workflows. One supports the review of capped-expiring trades and the preparation of pricing workbooks and customer communications, while others handle the production of fixed income rate lists, coverage dashboard preparation and trade confirmation reviews. Chatham professionals evaluate the AI-assisted output, refine it where necessary and decide what reaches the customer.

Chatham Onyx and model routing

Chatham Onyx places assets, debt and derivatives in a single environment where clients and advisors work from connected, governed data, with artificial intelligence capabilities that maintain traceability back to the underlying source. Codex is used throughout the Onyx development lifecycle, helping teams plan, build, test, document, and review software.

“Codex helps our teams turn product vision into business capabilities more quickly, while our standards for accuracy, safety and accountability remain the same,” said John DeGuenther, Chief Technology Officer.

The Onyx platform works on GPT-5.6 Sol, GPT-5.6 Terra, GPT-5.4 and GPT-4.1. Tasks such as simple analysis and non-production testing are directed to the most cost-effective models, while GPT-5.6 is used for complex tasks that need to maximize accuracy and value. An Onyx feature, ChatFIN, surfaces patterns in historical market data, gives users insight into their portfolios, and finds and links legal documents covering debts, derivatives and lease terms.

Expert capabilities and next steps

The case study states that Chatham uses OpenAI to perform structured comparisons, organize evidence and identify exceptions, and to help teams explore information more efficiently, and that as a result consultants can spend less time assembling information and more on interpreting it, managing difficult cases and advising clients.

Co-Chief Operating Officer Mike Noonan said the limiting factor for the company has been the time it takes for experts to reach the point where they can apply their expertise, and that OpenAI is helping to reclaim that time.

Chatham’s stated next steps are to expand commercial validation to other products, continue refining employee-created applications, and use Codex to help develop new features for Onyx.