Arabic-first.
Design and evaluate for the intended language experience.

Use this collection to review how language affects the whole task rather than only the text on a button. The software playbook covers direction, mixed-script content, formatting, and error recovery. The AI pilot guide adds bilingual evaluation, source review, and clear boundaries for model-assisted work. Test with the audience you actually intend to serve, using appropriate synthetic or approved material. These guides encourage deliberate review and a clear record of decisions instead of assuming that an English-language result transfers unchanged to an Arabic-first product.