Claude Gets Spreadsheets, Whiteboards and Dashboards: Anthropic Brings 9,000+ Apps Into the Chat

Working with AI used to look like this: you ask a chatbot to sketch out a formula, copy it into Excel, then move the result into a presentation. Now Anthropic is trying to eliminate that copying altogether. Claude has gained built-in spreadsheets, interactive whiteboards and dashboards, and the chat itself is turning into a workspace where everything is within reach.

What Exactly Was Added

The list from the announcement is short but telling:

• Sheets: spreadsheets, formulas and collaboration. This is a claim on the territory of Excel and Google Sheets.
• Whiteboards: interactive boards for brainstorming. This is already Miro's territory.
• Data: dashboards and analytics, in other words a lightweight replacement for part of BI tools.
• More than 9,000 connectors: from Salesforce to Adobe and Atlassian.

An important detail about the connectors: the apps work right inside the chat. This is not just "the model read data from a service" but the service's interface embedded in the conversation. Separately, Anthropic announced Meetings: the AI will be able to join calls, take notes and update documents. For now this is "coming soon", not "already available".

What It Looks Like in Practice

Imagine a product manager preparing a quarterly report. Previously the chain looked like this: an export from the CRM, pasting into a spreadsheet, manual formulas, a chart in a BI system, a screenshot on a board for discussion with the team. Each step lives in a separate service, and context is lost at every handoff.

In the new setup the same thing is done in one conversation. You ask Claude to pull deals from Salesforce, assemble a spreadsheet, build a dashboard and immediately open a board where the team will work through the reasons for the dip. The model remembers what happened in the previous step and doesn't require you to explain everything from scratch each time.

The main bet here isn't the spreadsheets themselves. It's unlikely that Sheets in Claude will catch up with Excel in depth of functionality any time soon, and Whiteboards won't replace Miro for those who live in it every day. The bet is elsewhere: the model sees data, discussion and documents at the same time and can act across them. Standalone SaaS tools lack this end-to-end context.

Limitations and Open Questions

For now this is an announcement, and much is unknown. From the post and the link to the press materials, it's unclear what the pricing is, what limits apply to spreadsheet size, whether Excel formats are supported, and how permissions are separated. The figure of 9,000+ speaks to the number of connectors, not their quality: some integrations will surely turn out to be shallow.

There is also the question of trust. When an AI gets access to a CRM, a task tracker and corporate documents, an error in a formula or an accidental edit becomes a real problem. Spreadsheets and dashboards built by the model still need to be checked by human eyes, especially if decisions are made based on them. And Meetings raises separate questions about the consent of call participants and the storage of recordings.

What to Do About It

Developers. If your service has an API, it's worth assessing whether it needs a connector in such a catalog. An interface inside the chat is becoming a new distribution channel, and anyone who isn't there risks dropping out of users' workflows.

Product managers. Take one recurring report, for example a weekly metrics summary, and run it through Claude as a pilot. Measure how much time it takes and how many errors have to be fixed by hand. Only after that is it worth discussing scaling.

Founders. This is a signal of pressure on simple horizontal tools: basic spreadsheets, lightweight boards, simple dashboards. The winners will be those with deep specialization, proprietary data and a built-in workflow that a single universal chat cannot reproduce.

The joke about a salary for Claude is funny, but behind it lies a serious idea: competition for the workplace is shifting from individual apps to whoever holds the context of the entire team.