How Endava builds an agentic organization with Codex

Source: OpenAI Date: 2026-05-28 URL: https://openai.com/index/endava

Summary

OpenAI case study (URL 403’d; title-only). Endava is a global IT services and software engineering firm (~11,000 people) that works as a delivery partner for enterprises. Their Codex deployment is framed as building an “agentic organization” — suggesting AI agents embedded across delivery workflows, not just individual developer tooling. IT services firms adopting agentic coding at this scale changes their economic model: fewer billable hours per deliverable, potentially higher margin if they can keep the delta.

Implications

• Enterprise distribution: IT services firms like Endava are a multiplier channel — their Codex adoption propagates into every client engagement they run, meaning one partnership announcement implies hundreds of downstream enterprise contexts.
• Agent-fleet operability: “agentic organization” as a framing unit (vs. individual developer productivity) is notable — it implies orchestration across teams and projects, not just per-seat tooling, which raises questions about how agent work is tracked, reviewed, and billed.
• Governance/constraint: services firms operating under client contracts have stricter IP and data-handling obligations than product companies — Endava’s adoption suggests OpenAI has cleared (or is clearing) enterprise data-boundary requirements that were previously blockers.