Henry Ford would have called much of what we're doing in legal tech right now "building a faster horse."

Jordan Furlong recently published a two-part series worth reading. His argument is simple: AI isn't going to replace lawyers, but it is going to replace most of what lawyers currently do. Furlong frames the split as two spheres:

Sphere 1 is legal production: the researching, drafting, summarizing, and analyzing that fills most of a lawyer's billable day.

Sphere 2 is legal judgment: the hard uncertain calls, the strategic counsel, the moments where a client can't move forward without someone who knows the road ahead.

Sphere 1 work is already migrating to machines. Not next decade. Now. The firms whose revenue depends on volume billing for production work are standing on ground that's eroding under their feet.

Where lawyers remain essential is at the inflection points. The hostile takeover that could collapse. The ruinous lawsuit where strategy has to be exactly right. Judgment, composure, and experience matter more than producing a well-formatted memo.

Here's where I want to push Furlong's argument further.

If the future lawyer is selling judgment rather than task output, the tech we build for law firms needs to support that shift, not simply accelerate the old model. A lot of legal tech right now is still optimized for Sphere 1. We're making lawyers faster at doing things clients will soon do for themselves.

Henry Ford would have called this building a faster horse.

The question is what Sphere 2 tooling actually looks like. Better risk sensing. Better pattern recognition across matters and industries. Better ways to synthesize messy, incomplete information into something a lawyer can act on quickly. We're not replacing judgment. We're giving lawyers better inputs for it.

Furlong's second article covers the business model implications. Post-AI firms won't look like traditional pyramids. You don't need layers of associates if machines handle production work. You need smaller teams of high-judgment lawyers supported by diverse professionals and backed by strong business intelligence.

The billable hour is a terrible metric for Sphere 2 work. The best strategic call of a client's life might take ten hours. Value-based pricing, subscriptions, success-based fees. These aren't radical ideas. They're the logical conclusion of a market where the work that matters can't be measured in time increments.

If you're building legal technology, it's worth asking honestly: am I making Sphere 1 faster, or am I making Sphere 2 possible? Both have value right now. Only one has a future in law firms.

Part 1: Part 1
Part 2: Part 2