Endava Redesigns Its Organization Around AI Agents; OpenAI Publishes Case Study
UK-based IT firm Endava (NYSE: DAVA) has redesigned its entire software delivery process around AI agents using OpenAI's ChatGPT Enterprise and Codex, cutting requirements analysis from "weeks to hours," according to a case study published by OpenAI. Released on May 28, 2026, the study frames Endava as having transitioned into an "agentic organization" and presents it as a flagship example of enterprise Codex adoption.
Endava is a global IT services company focused on software development for banking, insurance, and retail clients. It announced a strategic partnership with OpenAI in July 2024, rolling out ChatGPT Enterprise licenses to its more than 11,000 employees worldwide. Building on that foundation, the company has now expanded into using Codex—a cloud-executed AI coding and workflow agent—across requirements analysis, design, specification, development, and operations. Endava pairs Codex with its own AI-native delivery framework, Dava.Flow, codifying the judgment of senior engineers into agents to scale expertise across the organization.
One cited example is feeding the transcript of a two-hour meeting with the legal team into Codex to generate a requirements specification. Analysis that normally takes weeks can be compressed into hours, and design diagrams and specifications can even be generated in real time during client sessions (as detailed by Memeburn). By fundamentally reworking a traditional "human-centric sequential process" into an "agent-first parallel workflow," the case illustrates an evolution beyond mere coding assistance toward an organizational transformation tool.
Technically, Codex offers GitHub integration and cloud sandboxes, and controls agent behavior via an "AGENTS.md" file placed at the repository root that describes project goals, constraints, and conventions. Enterprise plans include governance features such as RBAC, OS-level sandboxing, approval gates, and audit logs—key for serving banking and insurance clients. Endava itself has published an article emphasizing the importance of "Day Two" (continuous optimization), signaling a focus on the post-deployment operational phase.
Competitors such as Anthropic's Claude Code and Cognition's Devin/Windsurf are discussed in parallel, but Endava stands out for deeply integrating ChatGPT Enterprise, Codex, and GitHub, and for pushing into the upstream requirements-definition stage. Reactions on X were largely favorable, with praise for cutting analysis from weeks to hours and the view that the shift to an "agentic organization" in organizational design is the real essence. From an e-commerce perspective, some noted that compressed development costs could reshape agency margin structures. Meanwhile, specific pricing and benchmark figures remain undisclosed (Enterprise is sales-inquiry based), and cautionary views about limits at large scale were not prominent within this scope.