Anthropic adds two tools to Claude: dashboards and animations in chat

Anthropic has released two new tools for Claude aimed at business users. One builds dashboards from real-time data. The other creates animation and other visual content. Both run inside the chat, with no export to Tableau, Power BI, or a video editor. The user describes the task in plain text, and the model returns a finished visual result.

What exactly is new

The announcement says little, and that matters. There are two tools. They target enterprise customers. The first handles on-the-fly data analysis and visualization; the second generates animated content. The source material does not give names, limits, pricing, or availability timelines for different plans.

• Number of tools: 2
• Audience: business users
• Interface: the Claude chat, no external services
• Tool 1: dashboards and real-time data analysis
• Tool 2: animation and visual content
• Pricing, limits, list of data sources: not disclosed

Where this came from

The logic is easy to trace. In June 2024, Anthropic launched Artifacts, a separate panel next to the chat where Claude displays code, pages, and graphics. In October 2024, the analysis tool appeared: the model wrote and executed JavaScript to parse CSV files. File creation, including spreadsheets and presentations, came later. The new tools look like the next step: less text in the response, more of a finished artifact you can work with.

Competitors are heading the same way. OpenAI has Canvas and data analysis in ChatGPT, and Google has Canvas in Gemini. The enterprise segment remains one of the key sources of revenue for all three, and a visual result is easier for a manager to grasp than a paragraph of numbers.

Where the pitfalls are

The phrase "in real time" needs clarification. It may mean a connection to live sources through connectors, or it may mean a quick recalculation of an uploaded file. For a company, the difference is fundamental. The first case requires access permissions, auditing, and an understanding of where the data goes. The second question is accuracy. A dashboard built by a language model looks convincing even when the aggregation is calculated wrongly. Checking formulas and queries remains the human's job.

Animation is simpler and more complicated at the same time. Most likely, this means programmatically generated graphics (code, SVG, HTML) rather than a video model on the level of Sora. The announcement does not specify this, so it is our editorial assessment, not a fact. For explanatory diagrams, process demonstrations, and presentations, this format is usually enough.

What to check in practice

If your team already pays for Claude for business, start with three steps. First, take a typical report, such as weekly sales, and manually compare Claude's dashboard against your existing BI system across 5–10 metrics. Second, ask your administrator which data sources are connected and what information leaves the corporate perimeter. Third, look at what you can do with the result. If it cannot be exported, embedded in other systems, and versioned, the tool will work for quick drafts and one-off reviews, but not for regular reporting.