Generating & Editing AI Text

What I do
Generate copy by querying generative text tools like ChatGPT, modifying queries to get better outcomes, and editing the results to quickly create copy for UX and other applications.

Case study: Creating notifications for an airline app
As a senior content designer / strategist for AKQA working to relaunch a major airline app’s Day of Travel experience, i used ChatGPT to create a wide variety of messages across user touch-points that were clear, concise, and varied.

Challenges
Writing dozens of messages on a limited budget/timeframe was only half the challenge: how do you vary up copy when you often need to notify a passenger about something several times (eg, asking them to check in or alerting them about a flight disruption) in an obtrusive way (eg, push notification). By using AI to generate a list of options in seconds, I had it operate as a junior copywriter, picking the language I thought was clearest and would resonate best, then asking it to extrapolate on that. The original language provided by the AI was also too perky and long, so I asked it to modify accordingly.

The Outcome
Notifications for the app were not only succinct and clear — using language that would be common to users — they were also varied, thanks to ChatGPT’s ability to quickly come up with 10 different ways to word one user action, eg, checking in or proceeding to a gate. This means users are less likely to find notifications to be redundant or annoying.

Related Work