@rjs True, attention spans are rock bottom. The writing angle is essentialy a substitue for the wireframes. Maybe we do just go back to creating wireframes. As you mentioned, whatever gets people dicussing anything other than hi-fidelty UI descions
@froessell Equally though, token architecture are where the foundations are built; brand, contrast ratios etc. they remove the mundane and allow focus on the overall experience
I do think a JTBD approach could help here though.
Not βhow do we add AI chat?ββ¨
But:
- what job is the customer actually trying to make progress on?
- are they exploring, deciding, narrowing, comparing, or resolving uncertainty?
- when does conversation help, and when does it interrupt?
The interesting opportunity probably isnβt replacing interfaces with chat. Itβs designing systems that understand when someone is in:
- browse mode
- search mode
- decision mode
- support mode
And adapting the interaction model accordingly, is my 2 pence
Can vs should. You can put a natural language bar in every app. But @bchesky is right that nobody's figured it out, because browsing and searching are different mental modes. People open Doordash to look at options, not to type a sentence. Giving someone a blank text field when they came to browse is just friction @benhylak
Can vs should. You can put a natural language bar in every app. But @bchesky is right that nobody's figured it out, because browsing and searching are different mental modes. People open Doordash to look at options, not to type a sentence. Giving someone a blank text field when they came to browse is just friction @benhylak