LinearB launches Agentic Studio for querying engineering data and building metrics

LinearB launched Agentic Studio on November 2 with a natural-language interface for querying software-delivery data, creating metrics and dashboards, and supplying engineering context to external AI clients. The company says the experience is generally available to every LinearB organization, while administrators control which user roles can access it.

Agentic Studio has two main interaction modes at launch. In Explore mode, users can ask questions about their organization’s engineering data. The system returns an existing LinearB Catalog metric, generates one when needed, or says when the available data cannot answer the question. In Build mode, users describe a metric and visualization in plain language. The system then creates or changes the metric and widget and runs the generated query before returning the result.

LinearB says each answer includes the chart, figures, widget and date range used as supporting context. The values come from live calls to the company’s public API rather than solely from the AI assistant. The dashboard library separates Personal, Shared and LinearB Catalog content. Users can search and favorite dashboards, share personal dashboards organization-wide as view-only, and clone dashboards.

The initial Catalog dashboards cover AI return on investment, engineering health and the software factory. LinearB says its AI dashboards track adoption across GitHub Copilot, Cursor, Claude Code, Amazon Q and Kiro, along with merge frequency and cycle time. Where providers expose the necessary data, the dashboards also track token use and estimated AI cost. Delivery views include cycle time, deployment frequency, mean time to restore, change failure rate, pickup time, review time, pull-request size and rework.

Company Admins can turn Agentic Studio on or off across an organization and grant access by Admin, Editor and Viewer roles. Users outside the selected roles remain in LinearB’s Classic interface. If LinearB AI Services is disabled, dashboards continue to work, but the AI Assistant and Build with AI do not. Users with access can also switch back to Classic, where some functions—including the Metrics Dashboard, custom dashboards and custom project-management metrics—remained at launch.

External AI clients can connect through LinearB’s Model Context Protocol server. Those clients can find and run Catalog widgets, query supported engineering and AI-impact metrics, and retrieve information about teams, users, pull requests, issues, incidents and connected development tools. Personal dashboards and private or shared custom metrics created in Agentic Studio are not available through the MCP widget tools. OAuth requests inherit the signed-in user’s identity, while API-key requests use organization-level authentication. The MCP server and in-product assistant predate Agentic Studio’s general release; LinearB made both available to all customers in September.

LinearB is pairing the product release with company-produced performance evidence, not independent tests of Agentic Studio. In a benchmark covering 2.7 million pull requests, 83,000 developers and 253 organizations, LinearB says the highest AI-usage group reached 2.3 times its June 2025 merge rate by May 2026. The company characterizes that relationship as correlation, not proof that AI caused the increase. It also says that among the top 10% of organizations in the benchmark, 79% of agent-opened pull requests merged within 30 days, compared with 92% of human-only pull requests.

A separate LinearB customer case study says Kraken used the platform and its MCP server to investigate review delays, reducing 90th-percentile review time from two days and 15 hours to one day and three hours over four months. That result was published by the vendor and has not been independently audited. LinearB’s launch materials do not state pricing or plan-level entitlements beyond organization-wide availability and the AI Services requirement for assistant and Build functions.