From asking to doing: How the world is putting ChatGPT to work

Source: OpenAI Date: 2026-08-06 URL: https://openai.com/index/how-the-world-is-putting-chatgpt-to-work

Summary

OpenAI published its first country-by-country usage data for ChatGPT’s 1B+ users: at work, people are more than twice as likely to use it to complete a task or create something (writing, coding, analysis) than they are outside work; adoption is narrowing between early-adopter markets and Latin America, Africa, and Oceania; and multimedia is the fastest-growing use case (7.8% of messages globally, over 10% in Brazil and Colombia). Nearly 80% of all conversations still fall under three buckets: Practical Guidance (28.3%), Writing (28.1%), and Seeking Information (21.3%).

Implications

Directly on agent-layer convergence — the “asking to doing” framing is OpenAI’s own name for the task-completion shift, published as data the same week Google shipped the same shift as product (Ask Maps ordering, Gemini trip agents). Also a cost/economy of AI data point at the demand side: if work usage skews task-completion 2x over casual asking, that’s the usage pattern that makes per-task cost (not per-query cost) the meaningful unit — reinforcing the token-cost-as-operating-cost framing already tracked. The adoption-gap-narrowing claim is worth watching as a counter to the bubble narrative circulating the same week (Zitron’s Microsoft-concentration piece) — broader geographic usage is evidence for genuine demand distribution, though usage volume and revenue concentration are different claims and shouldn’t be conflated.