October 2026 Summaries

3 posts from Lightdash

Filter
Month: Year:
Post Summaries Back to Blog
Claude Dashboards is a feature that converts plain-language questions into interactive, refreshable dashboards using connected data sources, with each chart linked to the underlying query and results for transparency. Users can connect directly to warehouses such as BigQuery, Snowflake, Databricks, Redshift, or ClickHouse, or use MCP connectors to analytics tools like Lightdash, which can preserve existing metric definitions, semantic-layer context, and user permissions. The workflow involves connecting a source, requesting a specific dashboard, reviewing chart-level lineage, and optionally publishing the finished dashboard back to the analytics platform for team access. The feature is positioned as a way for business users to quickly begin analysis, iteratively explore questions, validate supporting data, and refresh results as source data changes, although dashboard quality remains dependent on the available data and context.
Oct 09, 2026 875 words in the original blog post.
Lightdash has introduced Autopilot, an agent designed to maintain business intelligence projects as faster dashboard creation can lead to unused, outdated, broken, or poorly organized content. Running on a configurable schedule, Autopilot can retire stale charts and dashboards, repair broken charts, flag slow queries with pre-aggregation suggestions, surface popular but buried content, and identify assets owned by former employees. Users can limit it to reporting-only mode, define affected spaces and staleness rules, protect verified content, review actions and explanations in an activity log, and restore deleted assets from trash. Lightdash reports that Autopilot has performed more than 21,000 actions across 40 customer organizations, while Ollie said it retired over 400 stale assets in four months. The feature is generally available to Lightdash customers and supports organizations’ existing Anthropic or OpenAI models, including use of their own API keys.
Oct 06, 2026 626 words in the original blog post.
Lightdash enables teams to create, customize, and manage reusable chart types that align with company branding, with each visualization backed by downloadable source code and a manifest for local development. Users can generate charts through Chart Studio with an agent or edit code directly, then store projects in Git, review changes through pull requests, and publish versioned updates via the Lightdash CLI. Saved charts remain tied to their existing chart-type version until users explicitly review and adopt an update, helping prevent unintended dashboard changes. Custom chart types can also be transferred between staging and production alongside saved charts, while the Chart Library offers prebuilt options such as sparklines, heatmaps, and customizable Cartesian charts that teams can install or fork for independent customization.
Oct 01, 2026 596 words in the original blog post.