Undo Raises €31M for AI Software Debugging

Undo, a UK-based innovator in software reliability, has successfully closed a significant funding round, securing €31 million (approximately $37 million). This capital infusion is earmarked for enhancing its AI-driven root cause analysis capabilities and expanding its international footprint. The round was spearheaded by Elsewhere Partners, signaling strong investor confidence in Undo's unique approach to tackling the complexities of modern software development.

The investment arrives at a pivotal moment for the software engineering sector, as the integration of artificial intelligence introduces unprecedented levels of complexity. Greg Law, Founder and CEO of Undo, emphasized the company's strategic positioning, stating, "We've dedicated years to developing deterministic program recording technology that provides essential runtime visibility. This is now critical with the ascent of AI." He further elaborated that the new funding will accelerate the embedding of Undo's solutions into AI workflows, bolster commercial expansion, and solidify its role in an AI-centric engineering paradigm.

Undo's technology addresses a growing enterprise challenge: maintaining system stability and visibility as AI-generated or AI-assisted code proliferates. The company's platform provides AI agents with crucial runtime context, enabling them to accurately diagnose complex software issues within multifaceted systems. By capturing the complete execution history of code as it runs, Undo generates self-contained recordings that empower AI models to perform more effective root cause analysis, ensuring even AI-generated code remains understandable and manageable.

Market data indicates a robust trend of investment in the operational layers supporting AI-assisted software development. This includes areas like software supply chain security, AI agent governance, cloud operations, and developer tooling. Undo's focus on runtime context places it squarely within this vital segment, offering a solution to the reliability concerns arising from increased system intricacy driven by AI. Recent funding activities within the UK, involving companies like Cloudsmith, Geordie AI, Trent AI, CodeWords, Overmind, and Toyo, further highlight domestic momentum in securing AI-enabled software infrastructure.

Rod Favaron, Operating Partner at Elsewhere Partners, commented on the market need, noting, "AI is making code increasingly unmanageable, introducing elements that engineers struggle to comprehend, trust, or debug." He added, "Undo ensures engineering teams can effectively manage complex systems in an AI-first world and provides the essential runtime context for enterprise-grade, AI-assisted software." This sentiment is echoed by early adopters; Palo Alto Networks, through its Senior VP of Engineering Suresh Sangiah, highlighted Undo's ability to deliver crucial visibility, stating, "Undo provides the visibility needed to catch and correct errors before they become an operational problem for our customers, enabling automatic root cause analysis. Undo often autonomously finds the root causes in minutes."

Undo's internal benchmarks demonstrate a significant uplift in AI model performance when provided with runtime context. Without Undo, current AI models identify the root cause of complex bugs in only 38% of cases. However, with Undo's runtime context, this success rate jumps to 92%. Furthermore, when AI models can solve a problem without Undo, using its context results in fewer computational resources (tokens) being consumed. Customer feedback indicates that root cause analysis is completed up to 100 times faster with Undo compared to previous methods.

With this new funding, Undo plans to substantially scale its product development, customer support, and go-to-market teams across the United States and Europe. This expansion aims to meet the evolving demands of software engineering as AI integration continues to accelerate, ensuring Undo remains at the forefront of providing essential visibility and control in complex, AI-augmented development environments.