Vision
Vision
When Judgment Stops Being Scarce
A complaint response that used to take a law firm associate 16 hours now takes under four minutes. Harvard Law researchers documented 100x productivity gains on judgment-intensive tasks at major firms. Not a single one anticipates reducing attorney headcount.
Those two facts, sitting side by side, are where this gets interesting. Professional services have spent a century building intricate pyramids around one assumption: that structured discretion is scarce and expensive. Goldman Sachs deploying Claude for compliance work, and Garfield.Law winning approval to deliver legal services entirely through AI, suggest that assumption is quietly dissolving. What the firms do next reveals something about what "judgment" has been all along.
When Judgment Stops Being Scarce
A complaint response that used to take a law firm associate 16 hours now takes under four minutes. Harvard Law researchers documented 100x productivity gains on judgment-intensive tasks at major firms. Not a single one anticipates reducing attorney headcount.
Those two facts, sitting side by side, are where this gets interesting. Professional services have spent a century building intricate pyramids around one assumption: that structured discretion is scarce and expensive. Goldman Sachs deploying Claude for compliance work, and Garfield.Law winning approval to deliver legal services entirely through AI, suggest that assumption is quietly dissolving. What the firms do next reveals something about what "judgment" has been all along.
Proving Ground / New Model
Goldman's Compliance Bet Signals Where AI Agents Actually Are
Goldman Sachs chose compliance and accounting for its first major AI agent deployment—domains where high regulatory stakes, zero error tolerance, and complex judgment at scale create a threshold test. Success here signals something different than success in lower-stakes environments. The domain choice itself reveals where agents actually are on the capability curve.
Why Anthropic Engineers Spent Six Months Inside Goldman
Goldman Sachs didn't implement an AI platform. Anthropic engineers embedded at the bank for six months to co-develop systems. This deployment model exists because the gap between model capability and production reliability remains wide in high-stakes environments. Six months of embedded engineering to reach "launching soon" without a firm date—the approach itself signals where enterprise AI actually is.
The Goldman Deployment
Goldman Sachs deployed Claude across 12,000+ developers, reporting 30% faster client onboarding and 20% productivity gains. The numbers sound impressive until you ask what they measure.
"Developer productivity" typically tracks code generation speed in isolation, not whether the code ships faster or works better. A controlled trial found developers using AI tools took 19% longer to complete tasks while estimating they were 20% faster. Meanwhile, "30% faster onboarding" measures time-to-activate but ignores abandonment rates, error correction, or downstream service calls.
Goldman's CIO describes these agents as "early stages." They assist with coding and compliance work, but humans still define specifications and regulatory parameters. Speed without context is just noise.
Further Reading
