Reading Anthropic's Economic Index Report on When People Use Claude
Anthropic's June 2026 Economic Index report is subtitled "Cadences." Starting with this edition, Anthropic samples a slice of conversations every day, so it can now see what people use Claude for by hour and by day of the week. The report also covers what Claude produces and what a new survey found, but this post follows the rhythm part.
Start with the week.
The share of chat and Cowork conversations categorized as personal use spikes from around 35% on weekdays to just under 50% on weekends during the sample period. Outside the workweek, users' conversations shift from business correspondence, marketing copy, and slide decks to emotional support, medical questions, and investment advice.
"Chat and Cowork" here means conversations in those two tabs of the Claude app. Same tool, different questions once the day of the week changes. The report says the shift is largest in high-income countries, and that Claude Code and API traffic show a similar pattern, just from a lower baseline of personal use.
Claude Code on its own looks like this.
On weekends, the Claude Code usage clusters that fall the most include backend architecture, API debugging, and data storage. Those that increase the most include AI agent design, quant trading, and gaming.
Conversations about starting a business also peak on Saturday and Sunday across countries, while job-application activity drops along with other work. Looking back, I think I'm about the same. On weekdays I mostly build work-related things, and on weekends I tend to use it for trading or the business I'd like to start. That's close to the weekend user the report describes.
There's a rhythm within the day, too.
People ask for news at 7 a.m. local time. Business correspondence (e.g., email drafting) traces the arc of the workday, with a slight peak at 10–11 a.m. One of the biggest spikes is recipe requests, which are 2.3 times more frequent at 6 p.m. compared to the average.
Media recommendations cluster in the evening, and sleep advice peaks in the few hours before dawn. Work done on nights and weekends skews toward tasks from higher-wage occupations, though the report notes it can't conclusively identify the jobs of the people asking. Dates show up as well: on April 14, the day before the US tax filing deadline, tax-related conversations were eight times as common as on an average day in May, stayed about as high on the 15th, and dropped sharply on the 16th.
That's the part I found fun. AI use changes from person to person, and with the time and the day, and those patterns follow an ordinary day: news in the morning, recipes before dinner, sleep questions before dawn. You can see in the numbers how far AI has made its way into everyday life.
Anthropic closes by saying that measuring hour by hour shows how the cadences of daily life are etched into its usage logs, and that this opens avenues for future research.
