The Operating System Thesis

If you've led a law firm long enough, you've lived through the cycle: impressive demo, signed contract, underwhelming results. The conventional explanations — wrong vendor, immature tech, resistant lawyers — don't explain the consistency of the pattern. The real problem is architectural, and it starts with decisions the vendor made long before the sales team walked into the room.

The second is the all-or-nothing fallacy. Firms assume that improving operations means a massive, disruptive overhaul — ripping out every system and replacing it simultaneously. That assumption is both paralyzing and usually wrong. There's an old observation — often attributed to Bill Gates — that most people overestimate what they can accomplish in a day and underestimate what they can achieve in a year. Operational improvement works exactly this way. Most improvements are incremental and compounding. Connect two systems. Standardize one workflow. Structure one category of data. Each step reduces the integration tax and creates a foundation for the next. No single step feels revolutionary. But the firm that takes one step per month looks dramatically different a year later. That said, there are moments when the right move isn't incremental — when the accumulated debt in the firm's infrastructure is so structural that the honest answer is to replace the foundation rather than keep patching it. Knowing the difference between a problem that calls for iteration and one that calls for a fundamental shift is itself a leadership judgment. But the default assumption — that any meaningful change requires burning everything down overnight — is the one that keeps most firms from starting at all.

A letter arrives at a personal injury firm. What happens next — the scanning, uploading, renaming, re-keying, and notifying — involves every person on the team and not a single second of legal work. It is information passing through human relay stations because the firm's tools cannot pass it to each other. This is the integration tax, and most firms have no idea what they're paying.

Most law firms are treating AI adoption as a purchasing decision. The ones that will actually pull ahead are treating it as an infrastructure decision — and the gap between those two approaches is widening fast.