Operating models from real work.
Every note here comes out of a problem we actually solved for a professional firm. Not product copy. The system underneath: how AI goes wrong when you treat it as a tool you install, and the operating model that keeps it useful, attributable, and safe. Written so a managing partner can hold the firm to it.
New field notes, when there's something worth writing down. No noise.
RSS-
Anthropic says AI writes most of its code now. The next sentence is the one that matters.
Jul 1, 2026The company that makes the model just reported that AI writes most of its code, and that human review has become the bottleneck. For a firm holding client data, that bottleneck isn't a problem to engineer away. It's the product.
For Owners and partners of professional firms (law, accounting, engineering, advisory)
-
AI is already in your firm. The question is whether anyone's operating it.
Jun 23, 2026Every owner is asking whether their firm is falling behind on AI. It's the wrong question. AI is already inside the firm. The risk is that it got there without an operating model.
For Owners and partners of professional firms (law, accounting, engineering, advisory)
-
Putting AI inside your Clio practice without handing over a privileged file
Jun 20, 2026Why 'just connect AI to Clio' is a governance decision, not an integration, and the operating model that keeps the assistant fast, attributable, and incapable of leaking a privileged document.
For Law firms running their practice in Clio
-
AI search over your file shares, without leaking the folders it shouldn't see
Jun 18, 2026The two ways 'just point AI at the drive' goes wrong, why the index layer is the real asset, and the one rule that keeps an AI assistant from showing someone the HR folder.
For Engineering, architecture, accounting, and professional firms with file shares and an aging DMS