A TREND RESEARCH EXPERIMENT
A TREND RESEARCH EXPERIMENT
Milan Design Week: 04/20 – 04/26, 2026
A TREND RESEARCH EXPERIMENT
In 2011, Ken Jennings lost Jeopardy to a machine and wrote on his final card: I, for one, welcome our new computer overlords.
In 2016, AlphaGo played Move 37, a move no human would have chosen, that looked like a mistake, and won the game.
Both times, the question wasn't whether the machine could calculate faster. It was whether it could see differently.
FAITH AI is that question, in a different room.
Every day of Milan Design Week, a cultural strategist (a human) picks one piece of furniture that best represents the 2026 Zeitgeist. So does FAITH. Both write the analysis. You don't know which is which.
If AI picks better than a human, that's interesting. But the cultural strategist designed and trained it on their methodology and their voice.
We genuinely don't know who will perform better.
That's the whole point.
How AI TREND RESEARcher Works
Every decision we make is shaped by a clear sense of purpose. Our journey has been anything but ordinary. Through every step, we've focused on staying true to our values and making space for thoughtful, lasting work.
AI trend researcher runs two agents against each other.
The first, a cultural researcher, reads the day's pieces, picks one, and drafts the signal analysis.
The second, an editor, prompted as thirty years on a newspaper copy desk, reads it back, flags fabrications, cuts redundancy, demands rewrites. The researcher rewrites from the marks.
The bot lives in Slack, holds a conversation, pushes back under pressure, and remembers every pick it has made so it doesn't repeat itself.
For those who want the architecture:
Claude Agent SDK,
Airtable for pre-analysed piece data,
Exa for parallel cultural examples,
500 lines of Python
Field notes
Day 0: INTERVIEW WITH AN AI TREND RESEARCHER
The first thing we did: interviewed the AI about why it does what it does.
DAY 1: TBD
DAY 2: TBD
MILAN DESIGN WEEK 2026: SIGNAL DATABASE
(the database the AI trend researcher queries and analyzes)