Anthropic Economic Index Connector Answers AI Job Questions

Asking a Dataset Questions in Plain English

On July 22, 2026, Anthropic shipped a connector that plugs the Anthropic Economic Index directly into Claude. The Index is the company's ongoing public research dataset measuring how AI is actually being used across the economy, built from analysis of real Claude conversations with privacy protections applied. Until now, exploring it meant downloading data and doing your own analysis. Now you can just ask.

• The Anthropic Economic Index connector went live on July 22, 2026 and works with any Claude model
• Answers draw directly from the Index dataset rather than the model's general knowledge
• Setup takes about a minute — enable it from the connectors directory in claude.ai, nothing to install
• The underlying data remains freely available on Anthropic's site for independent analysis

What Can You Actually Ask It?

The example queries Anthropic highlights give a good sense of the granularity: which occupations use AI the most, what the most common ways people in a particular state use Claude are, what sorts of tasks teachers reach for AI to handle, and what kinds of work people are automating versus augmenting.

Those are the questions economists, journalists, workforce planners and curious professionals have been asking for two years and mostly answering with surveys or guesswork. The Index replaces projection with observation — it reports what people did, not what they said they might do.

The Caveat Anthropic Puts Up Front

The announcement is explicit that the data reflects patterns in Claude usage rather than the labour market as a whole. That is an important limitation and it is stated plainly rather than buried. Claude's user base skews toward certain industries, regions and job types, so the Index is a well-measured sample of one platform, not a census. Anthropic encourages users to review the underlying data and its documented limitations when interpreting results.

That framing is the right one. A dataset that tells you honestly what it cannot tell you is far more useful than one that doesn't.

Why Ship This as a Connector Rather Than a Dashboard?

Because the interesting questions are the ones nobody built a chart for. A dashboard answers the questions its designer anticipated; a connector answers whatever you think to ask, including cross-cuts that would have taken an afternoon of spreadsheet work. Pointing a model at a structured public dataset and letting people query it conversationally is one of the more genuinely useful applications of the connector pattern — the same architecture that has been quietly improving across the AI tooling landscape all year.

It also lowers the barrier considerably. A teacher wondering how peers use AI, or a small-business owner sizing up which tasks are worth automating, now gets a grounded answer without learning a data tool.

Part of a Broader Transparency Push

The connector follows the Index's regular reporting cadence — we covered the June 2026 Cadences report when it landed — and arrives in the same window as the company's $200M Economic Futures research fund. Publishing the data, funding independent research on it, and now making it directly queryable is a coherent posture: put the measurements in public and let other people check the work.

What to Watch Next

The things worth tracking are how often researchers cite Index-derived findings in published work, whether the connector's answers hold up against independent analysis of the same raw data, and whether other labs follow with comparable public usage datasets. Cross-platform comparison is where this gets genuinely powerful — a single vendor's usage patterns are interesting, but several overlapping datasets would start to look like real economic measurement.

Sources: Anthropic — The Anthropic Economic Index connector — July 22, 2026; The University Network — July 2026.