Skip to content
Nonprofit Solutions AIMission-first AI field lab

Nonprofit AI field guide

Reviewed July 21, 2026

Plan for accessibility and language

Treat access, comprehension, and language quality as design requirements, not final checks.

01

Include the people affected

Ask who will create, review, receive, or act on the output and what barriers they may experience. Include people who use assistive technology, communicate in languages other than English, have limited digital access, or rely on plain language.

Participation should influence the workflow, not only validate a finished decision. Pay attention to power: staff and community members need a safe way to disagree with a proposed use.

02

Test the rendered result

Review structure, reading order, headings, link purpose, alt text, contrast, keyboard use, captions, document tags, and plain language in the actual channel where people will receive the content.

For translation, use qualified human review appropriate to the consequence and community. Check names, idioms, cultural context, reading level, dialect, and instructions. Fluency does not prove accuracy or appropriateness.

03

Preserve a non-AI route

Do not make a person interact with an AI system to access essential information, request help, exercise a right, or participate in a program when an accessible human route is required.

Publish how people can ask for another format, correct a problem, reach a person, and understand whether AI influenced the content or interaction.

Monthly field note

Get practical guidance in your inbox.

Monthly field guidance for mission fit, data boundaries, human review, and bounded AI pilots.
Required