OpenAI's analysis of 800,000+ messages from U.S. ChatGPT users reveals that 16.8% of work-related messages and 43.5% of occupation-specific messages involve tasks from other occupations. This "task crossover" lets a small-business owner draft copy, review contracts, or analyze finances; a salesperson explore customer data; a marketer troubleshoot websites without developers.

Crossover rates vary: 77% for customer experience workers, 75% for designers, 69% for HR, 56% for legal, 53% for marketers. Marketing and engineering tasks travel farthest across occupations. Financial calculation and tech troubleshooting appear as top outside tasks across seven occupation groups. Designers import 35.2% outside tasks but export only 1.7%; engineers import 18.5% but export 7.4%; marketers both import (24.3%) and export (8.9%) heavily.

Business size matters: outside-task share drops from 18.9% in 2–5 seat workspaces to 16.3% in 100+ seat workspaces among average users. AI acts as a generalist tool where specialists are scarce. Usage patterns signal occupational shifts before job descriptions change.