Start with the problem, and the tool becomes very helpful

In a past life, I wrote a couple of books, so I'm very familiar with writer's block. That moment when you stare at a blank page, knowing the story needs to move forward, but having no idea where the plot should go next. I had a similar experience after installing Claude CoWork: I opened it up, ready to experiment… and then just stared at the prompt box. I even clicked the Ideas button, hoping for inspiration, but got nothing.

The problem wasn't the tool but that I didn't actually know what I wanted it to do. The tasks I was working on were already being handled by other tools:

1. Cursor was building a new webpage for our thought leadership section
2. The Claude extension was checking Mondrian blocks in a course we had just built
3. Gemini was extracting text from another course so we could create a shorter version

I had broken one of my own cardinal rules, and Chris Hornby's mantra: What problem are you trying to solve? I was about to use a tool simply for the sake of using a tool.

Of course, the moment I actually asked that question, my mind went straight to the problem I'd been avoiding for months: our chaotic image library. Right now, Claude CoWork is sorting through it, identifying which courses each image appears in, bundling them into sensible folders, and creating a visual brief for each one.

Start with the problem, and the tool suddenly becomes very helpful.