Claude Dashboards had a short life under that name. Anthropic launched it in beta on 8 October 2026, in the same announcement as Claude Motion. Thirty-three days later, on 10 November, it left beta as Claude Data. If you searched for Claude Dashboards this week and landed on a page titled Claude Data, you are in the right place: it is the same product, your existing dashboards still work, and it now does a good deal more than draw charts.
This explainer is for anyone who uses Claude at work, whatever the data. It covers what changed with the rename, how a dashboard is put together, and the features that came with general availability: certified queries, live sharing, data stories, design systems, caching and tracked changes. A short section near the end looks at e-commerce and Amazon data, which is what we work on.
The October beta did one thing. You asked a question about your company's data in plain language, and Claude built a dashboard that stayed current as the data changed. Anthropic described it as suited to quick, exploratory questions, such as how this week's signups compare with last month's.
Anthropic's explanation for the rename is that the product now covers more than dashboards. Three things are different since 10 November:
Claude Motion, the animated-explainer tool that shared the October announcement, is a separate product and keeps its name.
You start in any chat by asking for what you need, or in the Artifacts tab by choosing Data and describing it. Anthropic's tip is to describe the question you want answered, not just the chart, and to name the table or the app if you know which one holds the answer.
Claude then writes a SQL query for each chart and runs it against your data source. The part that sets this apart from a chart pasted into a chat is what you can inspect afterwards. Every chart shows its query and when its data was last refreshed. You can click through to see what a number counts, ask Claude to explain it, ask for changes in chat, or edit titles, captions and charts by hand.
Claude Data needs a data connector. Anthropic names five warehouses in its Help Center: Amazon Redshift, BigQuery, ClickHouse, Databricks and Snowflake. Apps you have connected to Claude can be sources as well, with Salesforce opportunities as the example given. Without a connected source there is nothing to query, so this is the first thing to set up.
Every company that has more than one dashboard has the same problem: two pages show two different numbers for revenue, and both are defensible. Certified queries are Claude Data's answer.
A certified query is one your data team has approved as the definition everyone should use. Claude can reuse it when it builds new charts, and charts that rest on it show a certified badge. According to Anthropic's Help Center, anyone with write access to a Data artifact can certify: open the chart's query, edit it, turn on "Certify query", add instructions for when it should be used, test it and publish. Definitions can also come from tools where teams already keep them, with dbt, Looker and Cube named, through existing MCP connectors. Organisations that want review can keep certified queries in a Git-backed plugin and manage them through pull requests.
Two details matter in practice. Editing a certified query removes its certification, and dashboards that depend on it lose the badge until someone certifies the new version. And the badge is Claude's label for "a person on this team approved this query". It says nothing about the data source, and it is not an endorsement by Anthropic.
A dashboard starts private. You open it and click Share, and it can also be shared outside your organisation.
What a viewer sees depends on one setting. Without live sharing, a shared dashboard loads its data through each viewer's own connections. A colleague who has no access to the warehouse sees the layout and no numbers. With live sharing, which is in beta, people in your organisation see results from your access, including saved results.
Anthropic built three limits into it:
The warning from the Help Center is worth repeating in plain words: live sharing runs on your access, so anyone you share with sees those results even if they could never open the source. Read the page as a stranger would before you share it. People with view-only access cannot edit a dashboard or restore a version.
A dashboard answers "what are the numbers". A data story answers "what happened". Claude Data can draft a written story or a short narrated video from your metrics, and you review it before it goes to anyone. The Help Center describes several shapes this can take, from a scrolling story to a short video to a written analysis. It is the feature behind the rename, and the one most likely to replace the Monday slide that somebody rebuilds by hand each week.
Design systems. A dashboard can follow your organisation's design system, so it looks like the rest of what your team ships instead of like a default template.
Caching and refresh. Results are cached so pages load faster, and you set how often each dashboard should refresh. The mechanics are worth knowing: Claude refreshes data when someone opens a dashboard and the data is due. It does not run on a fixed schedule in the background. A Refresh data button re-queries on request.
Tracked changes. Claude Data records who edited what. You can preview an earlier version, which is read-only, and restore it. A restore applies for everyone and shows up as its own entry in the history. Edits you make by hand are recorded there too.
It is not a replacement for a BI platform, and Anthropic does not present it as one. The original pitch was quick, exploratory questions, and the product connects outward: the October announcement listed Amplitude, Grafana, Hex, Mixpanel, Omni, PostHog and Sigma among the tools a dashboard can be sent to, with more marked as coming. If your company runs governed reporting in one of those, Claude Data is the fast first draft and the place where a question gets asked, not the system of record.
It is also not a spreadsheet. Anthropic shipped Claude Sheets the same day for work that belongs in rows and formulas, and Claude Whiteboards for work that belongs on a shared board. We cover how the three fit together for one industry in Claude Data, Sheets and Whiteboards: what changed for Amazon sellers.
Online sellers are a good test of any dashboard tool, because their data is scattered. An Amazon business has sales in Seller Central, purchase orders in Vendor Central and spend in Amazon Ads, each with its own reports and its own idea of a day. None of those is a warehouse, and Amazon is not among the sources Anthropic names, so the data has to reach Claude Data through a connector or through a warehouse that already holds it.
That is the layer DataDoe provides: Seller Central, Vendor Central and Amazon Ads in one schema, 130+ documented tables, across all Amazon marketplaces, fetched daily at 2:00 in the seller's local time with some tables updated during the day. It reaches Claude in two ways. One is a custom connector, added under Customize → Connectors → Add. The other is a BigQuery listing that puts the same tables into a read-only dataset in your own Google Cloud project, and BigQuery is one of Claude Data's named sources.
Both routes are laid out on our Claude Data for Amazon page, and the practical walk-through is in how to build an Amazon sales dashboard in Claude Data.
Only the name. Claude Dashboards became Claude Data on 10 November 2026, and dashboards made during the beta keep working.
It is included in paid Claude plans and draws on the same usage limits as the rest of your plan. Anthropic's announcement lists paid plans only.
No. It refreshes when someone opens a dashboard whose data is due, at the cadence set for that dashboard, or when someone presses Refresh data.
Only if you turn on live sharing for that source. Then people in your organisation see the results of the dashboard's approved queries on your access, and nothing else.
Claude Data answers questions about your data with dashboards and stories. Claude Motion builds short animated explainers from a prompt and is a separate tool, in beta on Team and Enterprise plans when it was announced.
Anyone with write access to the Data artifact, according to Anthropic's Help Center. Teams that want stricter control can manage certified queries in a Git-backed plugin with pull request review.
DataDoe is independent of Anthropic and is not endorsed by it. This explainer is based on Anthropic's announcements of 8 October and 10 November 2026 and its Help Center as read on 11 November 2026. Certified queries and live sharing are in beta, and details may change.
.svg)
