Anthropic adds a Claude connector for exploring AI workforce data

Anthropic has turned its Economic Index into a Claude connector, giving users a direct way to question the company’s dataset on how people use AI at work. The feature was announced on July 22, 2026, and is available in claude.ai through the connectors directory.

The move matters because the Economic Index has been a research resource more than a mainstream product. Anthropic says the index is built from patterns in Claude usage and is intended to help answer questions about which occupations use AI, what tasks are being automated, and how usage differs by field or location. With the connector enabled, users can ask Claude those questions in ordinary language and request the underlying data behind an answer.

What changed

Instead of downloading datasets or reading static reports, a user can now bring the index into a Claude conversation. Anthropic says setup takes about a minute, works with any Claude model, and does not require installing separate software. Example prompts listed by the company include asking which occupations use AI most, how teachers use Claude, or how automation patterns have changed over the past year.

The connector is also framed as a guardrail against overreading the data. Anthropic notes that the index reflects Claude usage patterns, not the labor market as a whole. In practice, that distinction is important: the tool can surface trends inside Anthropic’s own product ecosystem, but it should not be treated as a complete census of how workers across the economy use AI.

Why it matters

AI labor-impact debates often rely on surveys, anecdotes, or broad forecasts. A conversational interface for real usage data could make that evidence easier for researchers, journalists, policymakers, educators, and workers to inspect. It may also help nontechnical users test narrower questions, such as whether a specific job family is using AI for drafting, coding, analysis, tutoring, or administrative work.

Anthropic says the full datasets remain freely available on its website, so the connector appears to be an access layer rather than a replacement for independent analysis. The strongest use case is likely exploratory: using Claude to locate patterns, then checking the source data and limitations before drawing broader conclusions about automation or job change.

Sources

• www.anthropic.com

Cover photo by Jakub Zerdzicki on Pexels, used under the Pexels License.