# Census BTOS supplement: writing is the leading generative-AI task reported inside US firms, and 65% of firms limit AI use to three tasks or fewer

Page: https://flyvolo.ai/en/changes/ev-20260401-technical-writer-1
Date: 2026-04-01 · Stage: Worker adoption
Can this record move a task judgement? only together with another independent record
Occupation: https://flyvolo.ai/en/careers/technical-writer

## Scope

Nationally representative, reference period November 2025 to January 2026. This is the worker-task layer of the same supplement rather than its firm-adoption layer: 23% of firms (41% employment-weighted) report workers using AI in work-related tasks, and among the tasks reported, writing, document analysis and information search lead. It is a count of firms reporting a task, not a measure of how much writing is done that way, and it does not distinguish technical documentation from any other writing — nobody asked what was written. Its value here is what it is not: unlike vendor telemetry, this is a statistics agency counting the same channel with no stake in the answer, which makes it the better source for worker adoption where it exists.

## Source

- U.S. Census Bureau, Center for Economic Studies (CES-WP-26-25), worker-task findings — https://www2.census.gov/library/working-papers/2026/adrm/ces/CES-WP-26-25.pdf#worker-tasks (primary source)

## What this means

A statistics agency asking firms what their workers actually do with AI reports writing at the top of the list, and it reports it with no stake in the answer. That matters more than the ranking itself: the same channel is visible in vendor telemetry, and this is the version of it collected by a party that is not selling anything.

## What it does not show yet

It counts firms reporting a task, not how much of any task is done that way, and it does not distinguish technical documentation from email, marketing copy or anything else written at work — nobody asked what was written. The 65% limiting AI to three tasks or fewer also cuts against the broadest readings: this is narrow use inside most firms rather than general adoption.

## How to verify it yourself

List the document types your team publishes and mark which have a machine-readable source of truth. Those are the ones this finding is about; the rest are not, and the difference is your job.


More: https://flyvolo.ai/llms.txt