Lama AssafStrategist, designer, builder.
Finding the simplicity in the complexity.
tldr
i like putting things together to see what they do. now and then the result surprises me, and i grin.
that’s what keeps me designing how people and AI work together.
i make tools for the choices that repeat, and keep the logic and the connections for myself. people get my time. AI gets the rest.
also: podcasts, melodic techno, tennis when i can get a game, salad from my own garden. (the pesto is worth asking about)

The airport, in short
I was hired to consult for a startup in a very complex topic: airports. Every board member gets a lot of reports, and decisions got lost. No one knows what they want, so I ran a two-hour workshop on a shared board. A swarm of agents, each with a persona, stress-tested the concept. The deck wasn’t enough, so I built a prototype with tutorials. They used it in all their presentations.
The sports startup, in short
A sports startup had an idea and a World Cup date. The first version served fans, analysts and investors at once, and delighted nobody. Ten AI reviewers were asked one question: how does this die? We chose one lane and said who it wasn’t for. A working prototype took eight weeks. What stopped it was on none of the lists: a regional situation, just as funding was coming.
The Winner’s Circle, in short
-
horse names are weirder than token names.
so strange names were never the problem for racing fans. crypto was. most had never held any.
-
we put the game first.
fans play for stars, and stars turn into $HRSE.
-
fans shouldn’t need a wallet to play.
-
we were a small team.
we ran the socials and the work on the ground too.
-
i led the team from the first idea to launch.
how it plays, what it rewards, who makes which part. when the budget got tight, the site became mine too.
-
the project that got me into horse racing.
a beast of a project, no pun intended.
also: ava, the platform’s first agentic prediction agent. 70 percent of its predictions came true.
What I’m thinking about
Research and Insights
When a product took two years, someone hired a designer. When it takes two days, nobody does.
The making got easier. The thinking got skipped.
An idea gets ignored for two reasons. It’s too early, or it’s told badly.
The first is timing. The second is my job.
There is no neutral voice. Robotic is a personality too.
Not choosing is still a choice. Just not yours.
The loudest complaint is often the one you can’t fix.
The rare request is where the product is. Count less, read more.
Talk
Designing for the future of AI products
Maggie Appleton · Dive Club
Talk
The design process is dead
Jenny Wen · Lenny’s Podcast
Talk
From SaaS to agentic UX
Damian Gribben · Belfast Design
Talk
Design systems in the age of agents
Cassie Groos · Hatch Conference
07 · The work