About The Context Window

I think about one question: what actually changes inside a company after it buys AI?

Most AI coverage is product announcements, benchmarks, and demos. Almost none of it survives contact with a real organization, where the work lives in workarounds and tribal knowledge, inside processes nobody has examined since the last transformation program. The Context Window is about that layer: how work actually moves through an organization, and what it takes for humans and agents to run it together.

I'm JD Fetterly. I lead AI transformation for strategic accounts at Within AI, working with executives to see how their work really happens, then find where agents accelerate the strategy they already have. Before this I spent over two decades on the buying side of the same problem — transformation, ERP, and data governance roles at Google, Apple, and GE Digital. I've sat through the vendor pitches. I've inherited the stale process maps. I know what happens after the ink dries.

The claim I keep testing here: AI isn't the hard part. Transformation is. The organizations that win won't be the ones with the best model access. They'll be the ones that made the structure of their work visible enough to hand parts of it over, governed by real operating contracts between humans and agents — who owns what, what "done" means, who's accountable when an agent gets it wrong.

My own team at Within runs this way. Agents handle most of the execution; people supervise for direction, judgment, and quality. Nobody lost a job — the reclaimed hours went to customers. Writing in public is how I pressure-test what I'm learning before I stake a customer conversation on it.

Where to start: