Undo lands $37M to give AI agents the runtime context to fix bugs

2026-06-15 · Source: AI – SiliconANGLE · Field: Technology & Digital — Software Development & Engineering, Artificial Intelligence & Machine Learning · Depth: Intermediate, quick

Summary

UK-based software debugging startup Undo Ltd. secured \$37 million in new funding, led by Elsewhere Partners, to expand internationally and integrate its technology with AI-powered code fixing tools. Founded in 2005, Undo's core offering records the full execution history of a running program, saving its runtime behavior to a single file, a distinct approach from traditional debuggers that primarily show code. This capability is increasingly critical as AI coding assistants generate vast amounts of code. Undo claims its runtime recordings significantly enhance AI agents' ability to identify bug root causes, citing benchmarks where AI models improved from 38% to 92% accuracy with Undo's data. Customers, like Palo Alto Networks, report up to 100 times faster root-cause analysis, highlighting the technology's value in addressing complex runtime bugs. The investment will support hiring across product, support, and sales teams in the United States and Europe.

Key takeaway

For AI Engineers developing or integrating AI coding assistants, consider incorporating runtime execution recording tools like Undo. Your AI agents will significantly improve their ability to diagnose complex bugs by analyzing actual program behavior, not just static code. This approach can boost root-cause identification accuracy from 38% to 92%. It also accelerates analysis by up to 100 times, making your debugging workflows far more efficient and reliable.

Key insights

AI agents require runtime execution history, not just code, to reliably diagnose and fix complex software bugs.

Principles

• Runtime context is crucial for AI bug diagnosis.
• Traditional code-only debugging limits AI effectiveness.
• Full execution history improves AI root-cause analysis.

Method

Record a program's complete execution history to a single file, then feed this runtime behavior data to AI agents for precise bug root-cause identification.

In practice

• Integrate runtime recorders with AI coding assistants.
• Use execution history to reduce AI token usage.
• Accelerate root-cause analysis by 100x with runtime data.

Topics

• Software Debugging
• AI Agents
• Runtime Analysis
• Code Generation
• Root Cause Analysis
• Venture Capital

Best for: Machine Learning Engineer, Investor, CTO, AI Engineer, Software Engineer, Director of AI/ML

Related on AIssential

• Benchmarking AI Agents on Kubernetes — AI coding agents struggle with system-wide impact analysis, often providing incomplete fixes despite effective code retrieval.
• Getting Better at Working With You: Compiling User Corrections into Runtime Enforcement for Coding Agents — Compiling user corrections into runtime enforcement rules significantly improves coding agents' compliance with preferences across sessions.
• Automated Root-Cause Subclassification and No-Code Fix Generation for Invalid Bug Reports — Automating invalid bug report subclassification and no-code fix generation significantly reduces software maintenance overhead.
• not much happened today — Application-layer AI products can achieve high value and market fit without proprietary foundational models.
• New Guide: 6 Mistakes Breaking Your Production Agents — Six common mistakes consistently break AI agents in production, but they are predictable and fixable.
• Handoff Debt: The Rediscovery Cost When Coding Agents Take Over Interrupted Tasks — Handoff debt, the rediscovery cost for coding agents taking over interrupted tasks, is significantly reduced by context-bearing handoffs.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI – SiliconANGLE.