After a quarter-century in the IT services market, Endava has come to a realization: the traditional Software Development Life Cycle (SDLC) has become a monument to its own inertia. Today’s challenge isn't how fast a programmer can type, but the fact that human engineers become a bottleneck if they aren't supported by autonomous systems. The company’s CTO, Matt Cloke, states it plainly: the survival of the service business now depends on becoming "AI-native." This isn't about the cosmetic installation of a chatbot for tips; it’s about a radical demolition of workflows where an AI agent is appointed as the primary link in solving any task.
Shifting to the DavaFlow Agent Methodology
Endava's transformation has vividly highlighted an old industry ailment: accelerating coding immediately exposes congestion in adjacent departments. As soon as engineers picked up speed thanks to AI, it became obvious that requirements gathering and business analysis were hopelessly lagging. In response, the company rolled out DavaFlow—a proprietary delivery methodology designed around autonomous agents. Now, OpenAI technologies are embedded into every stage: from brief preparation to final deployment. Essentially, this is a shift from a "human plus assistant" model to full-scale orchestration, where the agent handles the drudgery of planning and communications.
"If I don't have an agent running in the background, I feel like I'm wasting my time," says Matt Cloke.
This approach derivatively changes team architecture. Instead of the usual relay race, where a task lazily migrates from analyst to developer, Endava uses a combination of ChatGPT Enterprise and Codex to manage complexity in enterprise projects. Agents don't just write snippets of code; they automate the asynchronous work of team leads and coordinate processes. According to Cloke, there isn't a single segment left in DavaFlow where algorithms aren't on duty.
Economics and Culture Beyond the Code
The economic impact quickly leaked beyond IT departments. Endava’s legal teams now use AI to sift through documents, while project managers have offloaded progress summarization and reporting to Codex. For a company with 11,000 employees, this isn't just a software update—it's a hard pivot in leadership competencies. AI adoption here is viewed as a behavioral change: management is expected to get their hands dirty with the tools to push the technology deep into the organization.
The primary challenge now lies in the realm of accountability. Endava is consciously creating space for experimentation, acknowledging that friction will always exist at the interface between agents and engineers in critical systems. However, the company is betting on deep orchestration, where human expertise is required only for the final verification of what the autonomous system has built. In the current climate, this looks like an attempt to turn a boutique service shop into a scalable factory, where margins grow not by hiring more heads, but through the density of agent coverage.
