Moving from AI adoption to an AI operating model: How Docusign engineering went beyond tool rollout

Docusign, a leading electronic signature platform, has successfully transitioned from merely adopting AI coding tools to developing a comprehensive AI operating model across its engineering teams. The company's engineering leaders recognized that while initial enthusiasm for AI was high, early gains in productivity plateaued, leading them to focus on sustained adoption and integration of AI into day-to-day workflows.

To achieve this, Docusign established an Eng AI Champions team, comprised of six engineers, tasked with developing guardrails, sharing best practices, and conducting workshops. This central team then extended its efforts through AI Multipliers embedded in individual development teams. By the end of 2025, 93% of engineers were actively using AI coding tools, and lines of code updated per developer increased by approximately 52% compared to the company's starting point.

Key to Docusign's success was the concept of "teaching AI to speak Docusign." The company built specialized internal tools, such as Kody for coding work, Chronicle for observability, and DocuWiki for internal documentation. These tools, combined with AI-assisted code review (Krew), helped engineers work more efficiently by identifying potential architectural and security issues, generating proposed fixes, and integrating them into pull requests.

The company's approach emphasizes a feedback loop, where AI agents learn from the successes and failures of individual teams. By making these agents reusable across teams and visible to all, Docusign has created a scalable model for AI integration that continues to evolve and improve. Docusign's leadership believes that the combination of central coordination, feedback mechanisms, and reusable AI agents has enabled the company to sustain and expand its AI operating model, setting it apart from other AI initiatives that may have faltered.