A thing I do a lot these days is this: when I spot a shiny new tool on X, I copy it and chat with @dhh through a gem I made at Google, @GeminiApp. DHH reminds me who I am. A digital-native person trying to code with AI.
The most important component of writing clearly is simply to have high standards for clarity. Then if you write something unclear, you notice, and ask: what did I mean to say? You can just keep doing this over and over. And if you have high standards for clarity, you will.
Lighting product data is complex. Anyone who's tried to manage it knows that already.
Add 40 years of product backlog, what Colin Hough at B-K Lighting calls a living organism, and you stop having a clean PIM project. You have something messier than that.
Okuldan artan zamanlarımda, Anadolu'da köy köy gezerek halk müzikleri kaydeden sıradan bir müzik öğretmeniyim. Kayıtlarımı tüm dünya tanısın istiyorum. Sizden ricam bunu retweetlemeniz. Ayrıca kanalıma abone olmanız için link:https://t.co/uGfixT0jMs
Thursday, May 21 · Zoom
7:00 PM Turkey (GMT+3)
6:00 PM Spain (GMT+2)
9:00 AM California, USA (GMT-7)
Hosted by Sales Layer PIM.
Register → https://t.co/hjb354mEFy
If you can't make it, register anyway. We'll send the recording.
AI on a lighting catalogue only works on top of a foundation. The foundation is structured product data — built in a PIM (Product Information Management).
Skip the foundation and the AI on top just produces wrong answers faster.
Marisa Back from Sales Layer PIM will run a demo of Agentic AI on representative lighting data.
Forty-five minutes. One PIM implementation story. One AI demo. No slide-deck energy.
A real data structure handles regional variation without a person copying fields by hand.
German datasheet gets German text. US datasheet gets 120V specs.
No spreadsheet heroics.
The market is rarely the problem.
Bad product data is.
Scaling a lighting brand globally is 10% shipping hardware and 90% managing data.
Exporting to the US sounds simple. Then the catalogue has to carry 120V specs, Imperial units, and a second language layer.
Do that in spreadsheets, and time to market turns into a delay.
Get these pillars right, and timelines and budgets stop being guesses and become commitments. You move from “storing data” to building a system that actually runs the business.
How do I start a PIM (Product Information Management) project?
Starting a PIM (project is often treated like a software purchase. It isn’t. It’s an architectural shift.