Article
I think AI should be boring.
When I look for where AI might help a business, my favourite starting point is organizing all the things people hate doing. The things that take a bunch of time. That's where I want to start looking for a problem worth fixing.
It's not sexy. It doesn't begin with a new model or another tool someone is trying to sell you.
I often see businesses starting at the other end. They find a tool, then try to figure out where they can fit it into the company. Before you go through that exercise, look at the work itself.
What are you trying to fix? Which part takes time? What do people keep having to do that they hate?
Get those things together before choosing another AI tool. Then check what you already have.
Check the connections before shopping
Look at the connectors available in your existing tools. A connector is a way for tools to work together. Before rushing into a new model or a new system for running AI, see whether something you already use might help with the problem you've identified.
I usually recommend starting with ChatGPT, Claude, or Copilot if you already have it. Those are starting points to consider, not a reason to buy all three.
The order is what I'd change: identify the work you want to fix, then see what your existing tools can do for it.
Start by writing down the tasks people hate and the ones that take a bunch of time. Pick a problem from that list, and check the tools and connections you already have against that specific job.