Anthropic's 'Cadences' Economic Index Links Its Survey to Real Claude Usage, Finding Heavier Automators More Optimistic About Their Jobs

Overview

Anthropic published a new installment of its Economic Index on June 26, 2026, titled Cadences, a report the company says began from a simple observation: “One year ago, most Claude usage took the form of a conversation between a user and an assistant.” The report’s most notable addition is methodological. For the first time, Anthropic links what people say they want from AI to what they actually do with it, reporting that “[we] link survey responses to usage data from mid-May to early June using privacy-preserving methods,” according to Anthropic.

The headline finding from that linkage is that the people delegating the most work to Claude are also the most optimistic about their own economic future. Anthropic reports that “people with a higher share of automated sessions feel more optimistic about the effect of AI on their job outcomes next year,” with the “largest effects on expectations about positive impacts on future pay and ability to find a job,” per Anthropic.

What We Know

The survey is small but tied to real behavior. Anthropic reports that “[our] final linked sample consists of about 9,700 survey respondents,” according to Anthropic. The independent write-up at Tony Reviews Things describes the same cohort as “roughly 9,700 Claude users whose responses were linked to their actual behavior through a privacy-preserving system.”

Users are split between wanting help and wanting to offload drudgery. Anthropic reports that “[over] half of survey respondents expressed some version of wanting to collaborate with AI on work that feels meaningful, of wanting their career to still matter, and/or hoping that new industries arise and create new job opportunities,” while “just over half of respondents hoped for AI automation—specifically of the tedious parts of their jobs—so they could have more free time and more space for meaning outside of work,” according to Anthropic.

The survey measured expectations across six dimensions. Anthropic grouped expected impacts into “pay, job security, ability to find a new job (economic dimensions) and meaning, autonomy, and human interaction (intrinsic dimensions),” per Anthropic. Along those dimensions, Tony Reviews Things summarizes that “users with higher automation shares reported more positive expected impacts from AI on their work.”

Heavy delegation did not correlate with feeling deskilled. Anthropic reports that “57%” of respondents reported “feeling like AI has made their skills more valuable” and “68%” reported “learning more with AI,” according to Anthropic. Tony Reviews Things notes that “heavy delegators report learning at the same rate as everyone else.”

Expectations for near-term capability are aggressive. Anthropic reports that “[close] to 6 in 10 respondents chose a higher band for next year than for today” when estimating how much of their work AI could do, and that “[over] a third expect AI to be able to do most or nearly all of their work tasks next year,” per Anthropic. The finding is echoed by ExplainX, which writes that “[over] one-third of surveyed users expect AI to do most or nearly all of their work tasks within 12 months.”

Concern concentrates on more junior workers. Anthropic reports that “[respondents] were especially worried about job loss for their junior colleagues, with over one third stating that the probability of a junior colleague losing their job in the next year was over 60%,” according to Anthropic. Tony Reviews Things records the same figure: “[over] a third rated the probability of a junior colleague losing their job in the next year as over 60%.”

How the Data Was Gathered

The usage side of the report draws on a defined sampling window. Anthropic reports that “[data] in this chapter cover chat and Cowork conversations sampled between April 10 and June 10, 2026,” and that the underlying survey was “launched in April 2026,” per Anthropic.

The company also describes three methodological changes in this installment. Anthropic says it now samples “data at a higher rate, allowing us to view usage patterns down to the hourly level,” introduces “a new classifier that labels the output of each conversation,” and shares “more granular data, breaking out results for chat and Cowork conversations,” according to Anthropic. ExplainX summarizes the same upgrades as a shift toward hourly telemetry, artifact classifiers, and a linked survey of about 9,700 users.

What We Don’t Know

The report does not present a single headline percentage declaring that automation has “overtaken” augmentation across all Claude traffic; its automation findings are expressed as correlations with survey sentiment rather than as one aggregate ratio.

The linked survey is also not a representative sample of the workforce. It consists of about 9,700 Claude users, according to Anthropic — people who already chose to use the product, and who then chose to respond. Optimism among heavy delegators may reflect who opts into heavy delegation as much as any causal effect of automation itself.

Finally, the survey captures expectations, not outcomes. The finding that close to 6 in 10 respondents expect AI to handle a higher share of their tasks next year, per Anthropic, is a forecast by users about their own jobs, not a measurement of tasks actually displaced.

Analysis

The throughline of Cadences is Anthropic’s own preferred framing: that heavier use of Claude has so far looked complementary rather than substitutive. Anthropic states that “[if] the human remained involved in the highest-value tasks, the pattern looks more labor-augmenting than labor-displacing,” according to Anthropic — a formulation Tony Reviews Things paraphrases as “more production from Claude does not mean less from the user.”

That framing is convenient for the company publishing it, and the report’s own numbers hold a tension the framing does not fully resolve: the same respondents who feel optimistic about their own prospects also, in large numbers, expect AI to do most of their tasks soon and worry that junior colleagues will lose their jobs. The optimism is real in the data, but it sits alongside an expectation of rapid capability growth that is harder to square with a purely augmenting future. Whether the augmentation pattern persists as models take on more of the work is precisely the question the Economic Index will have to keep measuring.