Endava's Agentic Organization Strategy with OpenAI Codex

Endava, a global software contracting firm, has implemented OpenAI Codex to transition into an "agentic organization." By codifying senior expertise into AI agents, Endava has shifted its operational model from manual code production to the oversight of AI-generated work, resulting in an exponential increase in output quality and the ability for small teams to deliver high value in condensed timeframes.

Codifying Senior Expertise for Knowledge Transfer

Codex serves as a mechanism for scaling senior architectural judgment across an organization. Senior architects can articulate their specific points of view and best practices, which Codex then uses to guide junior developers in real time. This transforms the traditional mentorship model, where senior guidance is typically delivered through slow processes like pairing and code reviews, into a scalable asset that can guide multiple teams in parallel.

According to Mike Krolnik, Global SVP of Agentic Architecture at Endava:

"Senior architects like myself... are able to articulate what we want, and Codex makes that an accessible piece of information for the more junior people on the team. And from the junior perspective, they’re able to adopt this tool and create senior, mature-level outputs."

Compressing the Software Delivery Lifecycle

Endava has moved away from sequential hand-offs between analysis, design, and build stages, instead using Codex as a unified tool to handle these phases simultaneously. This integration has compressed workflows that previously took weeks into a matter of days.

Requirements Analysis and Specification

In one instance, Endava's engineering team handled a complex request from the legal department involving the review of thousands of contract pages. Rather than spending weeks on back-and-forth communication to define requirements, the team recorded a two-hour deep-dive meeting, fed the transcript to Codex, and generated a working requirements specification. This process reduced a potential two-week revision cycle to two one-hour meetings.

Real-time Client Collaboration

Codex is used during live client sessions to rapidly produce design documents, diagrams, and software architecture visualizations. This acceleration of the feedback loop allows teams to illustrate ideas and reach consensus with clients more quickly.

Strategic Implementation Guidelines

Endava identifies three core lessons for organizations looking to implement similar agentic workflows:

• Codify Senior Judgment: The highest leverage is achieved by capturing the expertise of senior architects so junior members receive high-level guidance during execution.
• Expand Beyond Coding: Codex should be treated as a general "desktop agent" rather than a simple coding assistant. Its value is maximized when applied to requirements analysis, design, client communication, and operations.
• Start with Non-Coding Workflows: For teams new to the tool, Endava recommends starting with requirements analysis or design documentation—areas where coding tools have not been previously used—to realize the fastest value.

As Mike Krolnik summarizes:

"Codex magnifies every skill I have, and everybody who learns how to use it gets every skill they have magnified."