What Just Changed
On July 22, 2026, Anthropic launched the Anthropic Economic Index connector for Claude. This new tool lets anyone ask questions about AI and work directly in claude.ai, using natural language. Instead of downloading datasets or writing SQL, you can simply type questions like “Which occupations use AI the most?” or “What sorts of tasks do teachers use Claude for?” and get answers grounded in the Index’s data.
The Anthropic Economic Index itself has been around for a while, providing data on how AI is actually used in the economy. It has been a resource for researchers, journalists, and policymakers. But the connector is a deliberate step to make that data accessible to everyone—not just experts. As Anthropic puts it, they want the Index to be “just as accessible to anyone curious about how AI fits into their field or day-to-day life.”
For product builders, this is more than a new query interface. It’s a case study in turning a public dataset into a conversational tool. The underlying philosophy is worth noting: data’s value isn’t just in its existence, but in how easily it can be explored.
How the Connector Works
Getting started takes about a minute. In claude.ai, you open the connectors menu, find the Anthropic Economic Index in the directory, and enable it. Once enabled, it works in any conversation with any Claude model—there’s nothing to install. From there, you ask questions the way you’d ask a colleague.
You can start broad, like “What does the Index say about my industry?” and then drill into specifics. If you want to see the underlying data behind any answer, you can ask Claude to show it to you. The connector is designed to be flexible, supporting both high-level summaries and deep dives.
Anthropic emphasizes that the Index reflects patterns in Claude usage, not the labor market as a whole. Claude will point you back to the source data and its limitations as you explore. This built-in honesty is a key design choice—it prevents users from overgeneralizing the results.
The example questions Anthropic highlights include:
- “Which occupations use AI the most?”
- “What are the most common ways people in Colorado use Claude?”
- “What sorts of tasks do teachers use Claude for?”
- “What kinds of tasks are people automating with AI? How has that changed over the past year?”
These show the range of queries you can make, from broad occupational trends to specific geographic or task-based questions.
Practical Use Cases and Implementation
For product builders, the connector offers several lessons. First, it demonstrates how to lower the barrier to data exploration. Users don’t need to learn SQL or data analysis—they just ask questions. This expands the audience from professional analysts to anyone curious about AI’s role in their field.
Second, it shows the value of integrating data tools into existing workflows. The connector lives inside claude.ai, so users don’t have to switch between a database and a chat interface. This reduces friction and makes it more likely people will actually use the data.
Third, it highlights the importance of building in limitations. Anthropic doesn’t hide the fact that the Index is based on Claude usage. Instead, Claude actively reminds users of this limitation. For product designers, this is a reminder: being transparent about data scope and bias builds trust, especially when data could be misused.
If you’re designing a similar data query feature, consider these questions:
- What questions will your users ask?
- Do they need to see raw data, or are summaries enough?
- How can you integrate the feature into your existing product to minimize friction?
Anthropic’s design lets users do both—get summaries and drill into raw data—which is a balance worth considering.
Limitations and Trade-offs
While the connector is powerful, it has limitations. Currently, it’s only available on claude.ai. If you want similar functionality in your own application via the API, you’ll need to integrate the dataset yourself. The good news is that the full datasets remain freely available on Anthropic’s website, so developers can download them and build custom tools.
Another limitation is that the Index is based on Claude usage data. It doesn’t represent all AI tools or the entire labor market. If you’re using this data for decision-making, you need to understand how it was collected. Anthropic’s announcement doesn’t detail the collection methodology, but it’s clear that the data reflects Claude’s usage patterns, not a comprehensive economic picture.
This means the answers you get are indicative of how people use Claude, not necessarily how AI is used across the economy. For example, if a certain occupation shows high Claude usage, it might be because that profession has adopted Claude specifically, not because AI in general is prevalent there.
For product builders, this is a cautionary tale: when you provide data to users, be honest about its scope and biases. Don’t pretend it’s more comprehensive than it is.
Concrete Takeaway
If you’re interested in AI’s impact on work, the connector is worth trying. It takes about a minute to enable, and you can ask questions about your own industry. This isn’t just about satisfying curiosity—it’s an opportunity to observe how Anthropic designs conversational data exploration.
For product developers, the connector offers a model for turning complex datasets into accessible, conversational experiences. Key takeaways include:
- Lower the barrier to entry: let users ask questions in plain language.
- Integrate with existing workflows: don’t force users to switch tools.
- Be transparent about limitations: build trust by acknowledging data scope.
- Offer both summaries and raw data access.
Finally, remember Anthropic’s reminder: the Index reflects Claude usage, not the entire economy. Treat it as a starting point, not the final answer. The connector is a step toward making data more accessible, but it’s up to you to interpret it critically.
Whether you’re a builder or a curious user, this tool is a practical example of how AI can help us understand AI’s role in work. Give it a try, and see what you learn.
Sources
AI-assisted summary compiled from the sources above, reviewed by a human before publishing.