You don't need to understand machine learning to use Foundry. Here's what happens, in plain English.
1. You have a computer. A Mac Studio — a powerful but normal-looking desktop computer that sits in your office. It plugs into a normal power socket and connects to your network. It's quiet, it's compact, and it doesn't need a server room.
2. We install Foundry on it. Foundry is the software that makes the computer able to do AI work. We come in (remotely or in person), configure everything, test it, and hand you a working system. You don't install anything yourself.
3. You give it work. Depending on what you're using it for: documents? Drop them in a folder. Support tickets? Connect your help desk. Internal knowledge? Point it at your shared drive. Code reviews? Connect your GitHub.
4. Foundry does the work. It reads, classifies, extracts, drafts, and organises — all on the Mac Studio in your office. No data goes anywhere.
5. You review and approve. Every output is a draft. A document extraction shows you the data for confirmation. A support response shows you the draft for approval. A code review shows you the issues to check. You're always in control.
6. It keeps running. Foundry monitors itself. If a model crashes, it restarts. If memory is getting tight, it warns you. If something needs attention, the dashboard tells you — you don't have to watch it.
The principle: The AI that touches your data runs locally. Everything else stays where it is.
Most teams are live within a week of the Fit Review.
The AI model runs on the Mac Studio. Processing happens there. Results stay there.
Foundry doesn't phone home to OpenAI, Google, or anyone else during processing.
If you feed Foundry a PDF, the original is kept untouched with a digital fingerprint. The extracted data is stored separately.
Foundry prepares drafts. A person reviews and approves before anything goes out.
You own the hardware. The Mac Studio is yours. Foundry is software on your machine. You can turn it off, disconnect it, or inspect it at any time.