Portfolio
Overloaded with emails? Not having the time to read through all documents or articles you'd like?
How can AI help users read smarter, not harder?
When thinking about how to solve that ambiguous problem, I first started brainstorming about what users would want from that feature: Do they only want a summary? Would they'd like the article to be read out loud in an accelerated way? Are they only looking for an answer to a specific question?
While exploring the set of features that would live in Help me read, I also explored different surfaces where that content could live on. We really were defining the feature as I was designing it.
Some early explorations
Side panel
Mainly targeting website content at first, I explored solutions living directly in Chrome, in its existing side panel.
Ephemeral dialog
I explored surfaces that could be quickly summoned by users and automatically disappear once their questions got answered.
Right click summaries
Right-click is a frequent action users perform when interacting with content. I explored solutions leveraging that pattern to show Help me read capabilities.
While the idea of showing off the full power of this feature on right-click wasn't scalable (for many reasons - too slow, too big and too costly mainly) we landed on using the contextual menu showing up on right click as an entry point.
With this approach, we were creating a cohesive way for users to think about what was available to them behind a right click: when performing a right click over a word, they would see Quick answers cards; over editable text, they'd see Help me write and when performing a right click anywhere else, they would see Help me Read.
Help me read uses Gemini AI to summarize content and lets you ask follow-up questions.
These capabilities are living in a new surface that can be opened via right-click and stays with users until they explicitly decide to close it.
