With Agentic Studio, LinearB provides the critical context needed to accelerate software delivery

LinearB Inc., an application development intelligence startup, has introduced Agentic Studio, a new offering aimed at tackling the congestion in software delivery pipelines caused by the adoption of AI coding tools. The startup claims that while writing code has become effortless with the help of generative AI tools like Claude Code and Codex, shipping that code into production remains as challenging as ever due to the larger size of AI-assisted pull requests.

According to LinearB's recent Software Engineering Benchmarks report, which analyzed over 2.7 million pull requests from 250 organizations, engineers relying on AI tools have seen a significant increase in merged code changes year-over-year. However, these AI-assisted changes are nearly 2.6 times larger, leading to longer review times and significant backlogs.

Traditional dashboards used by software engineers only provide metrics, lacking the context needed to resolve pull requests, resulting in a backlog that needs to be cleared. With Agentic Studio, LinearB aims to address this issue by normalizing data and insights from Git repositories and project tracking tools into a unified context layer.

This allows engineers to ask questions in plain language to understand the root cause of delays and receive recommendations to speed up the process. Co-founder and Chief Executive Ori Keren emphasizes that context has become crucial for both AI tools and human engineers. Agentic Studio is designed to assist both autonomous coding agents and human software engineers.

By integrating with protocols like the Model Context Protocol, it provides historical context such as rework rates, past production incidents, and code ownership to AI agents, preventing them from making unverified modifications to pull requests. Early adopters, like cryptocurrency exchange Kraken, have benefited from using Agentic Studio, cutting review times by over half within four months after identifying timezone misalignments as the main issue.

LinearB is also developing new agent skills to help AI agents operationalize the provided context. The first skill, an "effort skill," analyzes the complexity of a pull request, its proposed code changes, rework history, and risk to dynamically calibrate the computational effort required for coding agents to work on it. Future agent skills will focus on automated code review, merge governance, and token cost optimization.