Case study — Yahoo Mail
Telling people
when AI is reading
Disclosure, feedback, and settings copy for AI message summaries in Yahoo Mail.
In short: I wrote how Yahoo Mail discloses AI summaries: one clear disclosure, an honest accuracy note, and a feedback prompt that asks about the right thing.
The challenge
Honest about AI, without scaring people off
Summaries save time, but most people don't expect a personal email to be summarized by AI. The copy had to say clearly that AI is involved, admit it can get things wrong, and make it easy to turn off, without burying the summary under warnings.
The content challenge
Say it once, clearly
An early onboarding draft repeated "we're using AI" on screen after screen. That made the feature sound risky without telling people anything new. I pushed for one clear disclosure where it matters, with more detail one tap away.
The work
Writing the moments that matter
Settings description
Yahoo Mail uses AI to summarize the most important information from your email.
This technology is new and may make mistakes.
Names the technology, says what it does, and sets expectations in two short sentences.
Feedback prompt
Was this summary helpful?
The old prompt sat next to “Created by Yahoo Mail,” so it seemed to ask people to rate the label. Naming the summary makes clear what they're rating.
Summary card footer
Generated by Yahoo Scout. Mistakes are possible.
Short enough to sit under every summary without crowding it.
Outcome
What shipped
Consistent AI disclosure across settings, summaries, and feedback
A feedback prompt that asks about the summary itself
An easy way to turn summaries off, mentioned wherever they're explained