openaienmodel: gpt-5-mini-2025-08-07
How Endava is redesigning software delivery around AI agents
Key Points
- AI agents across delivery lifecycle
- DavaFlow — AI-native delivery methodology
- Leadership models AI use
Summary
Endava has redesigned its software delivery around AI agents by making OpenAI (ChatGPT Enterprise, Codex) the enterprise platform and introducing DavaFlow, an AI-native delivery methodology. AI agents are embedded across the delivery lifecycle — from meeting prep, requirements and discovery to engineering, deployment and operations — and adoption extends beyond developers into legal, finance, project management and commercial teams. The change focused on behavior and workflow redesign rather than just tooling, producing faster requirements, less manual reporting, and enabling non-engineers to build lightweight internal apps.
Key Points
- DavaFlow: an AI-native delivery methodology using OpenAI across planning, discovery, engineering, and deployment.
- Agents run continuously for meeting preparation, project summaries, inbox and communication automation, and stakeholder coordination.
- Codex + agents let teams generate governance reports, summarize progress, and build one-page apps, reducing reliance on spreadsheets and dedicated engineering effort.
- Outcomes: accelerated delivery, reduced manual coordination and reporting, and more internal tooling created by non-developers.
- Principles: treat adoption as behavior change (leaders must model use), create space for experimentation, involve non-technical teams early, and emphasize hands-on experience.
- Practical engineer actions: start using agents personally, embed agents in requirements and QA workflows, automate recurring reports, and measure cycle-time or lead-time improvements.
Next steps for teams
- Pilot agent-backed workflows in one part of the delivery lifecycle (requirements or release reporting).
- Track metrics (cycle time, coordination overhead) and iterate fast.
- Ensure leadership adoption to normalize agent-led behaviors.