Finally, Someone Put Actual Data Behind the AI Job Panic

Every week, I get another Slack from a product manager asking if AI is going to make our team obsolete. I point to the same three tweets. Nobody’s happy.

Enter the Anthropic Economic Index connector—a way to ask Claude questions about how people actually use AI in the wild, grounded in real usage data rather than vibes. The Index has existed for a while as a research dataset, but until now you had to download CSVs and run your own analysis. The connector makes it a chat interface.

Here’s the workflow: open the connectors menu in claude.ai, find the Anthropic Economic Index in the directory, and enable it. No install, no API key. Then you just ask questions like you’re talking to a co-worker who spent the weekend crunching numbers.

For engineers specifically, this is immediately useful. You can ask: “Which occupations use AI the most?” or drill into “What tasks are software developers automating with Claude?” or “How has that changed over the past year?” The answers come back with citations to the underlying data, which is still freely downloadable if you want to validate.

I tried it with “What are companies in the fintech space doing with Claude?” and got a breakdown of task categories—code generation, documentation, data extraction—with relative frequencies. That’s not just a curiosity. It tells me where competitors are reducing friction and where my team might want to focus effort.

Of course, the limitation is right there in the name: this is Claude usage data, not a census of the whole economy. The connector surfaces that caveat and links back to source documentation. But for making tactical decisions about automation priorities, it beats the hot take of the week.

For someone shipping code tomorrow, the practical takeaway is simple: instead of guessing which workflows are ripe for AI integration, ask the data. The connector takes maybe sixty seconds to set up and gives you something you can point to in a sprint planning meeting. That’s a big improvement over “I saw a LinkedIn post.”

Data won’t stop your PM from asking the question. But it might stop you from rolling your eyes.