# Now available: Sherlock can analyze Depot CI workflows and jobs (https://depot.dev/blog/now-available-sherlock-depot-ci-analysis) > By Pedro Guerra (Support Engineer at Depot) > Published 2026-05-27 We previously taught our AI assistant Sherlock how to analyze container builds, GitHub Actions jobs, CI analytics, and project settings. [Depot CI](/docs/ci/overview) now has the same dashboard-aware experience: when you're looking at a workflow or job, Sherlock can fetch the relevant context and help you figure out what happened without copying IDs, logs, or screenshots into chat. ## Why we built this Depot CI already gives you logs, CPU and memory metrics, job summaries, retry attempts, and a handy AI failure diagnosis in the dashboard. That context is useful, but debugging still often means moving between a workflow summary, a job page, or an external AI tool to get the information you need. Sherlock can help when you want an explanation that pulls those pieces together. It can start from the workflow or job you're already viewing and use the available Depot CI data to investigate. ## What Sherlock can analyze For Depot CI workflows, Sherlock can fetch run context, workflow metadata, other workflows from the same run, job statuses, timing, current attempts, and errors. For Depot CI jobs, Sherlock can fetch job metadata, workflow context, attempts, step details, job summaries, CPU and memory metrics, failure diagnosis summaries, and error log lines. When more detail is needed, Sherlock can pull logs for a specific step. That means you can ask questions like: * Why did this Depot CI workflow fail? * What changed between attempts? * Which step failed, and what should I try next? ## Failure diagnosis follow-up Depot CI failure diagnosis cards give you a short explanation of what went wrong and a suggested next step. When you want more detail, click **Ask Sherlock** from a failure diagnosis card.