Most AI features fail for
Most AI features fail for a boring reason: they live somewhere nobody
goes.
You ship an AI panel. It sits in a tab. The tab stays closed. The
feature quietly dies while the roadmap says it launched.
So I designed this one to have no home screen at all. It is a design from attio
It answers in Slack, where the deal conversation is already happening.
It writes to the CRM, where the record already lives. It runs in the
terminal, where the technical team already works.
Three surfaces. No new place to check. No new habit to build.
The design question wasn't "what should the AI look like." It was
"where does the work already happen, and can we meet it there instead
of asking people to move?"
That reframe kills a lot of beautiful dashboards. It should.
I'm a product designer working across marketing surfaces and complex
product systems. Open to product design roles — permanent or contract,
remote.
Founders: what percentage of your users have ever opened the AI
feature you shipped?
0 likes · 1 view · Sep 6, 2026
Most AI features fail for a boring reason: they live somewhere nobody goes. You ship an AI panel. It sits in a tab. The tab stays closed. The feature quietly...
