Why 90 days is the right cadence for AI delivery
Three-month engagement cycles retire risk faster than any other rhythm we have tried. Here is why that shape holds up even as projects get bigger.
ExploreInsights from the field
Field notes
What actually works when shipping production AI for mid-market law firms: evaluation rigor, model swaps, integration brittleness, and the lessons we paid full price for.
Building RAG that actually works in production
Most retrieval-augmented generation systems look great in demos and fall apart in production. Here are the decisions that separate the two.
ExploreWhat the AI hype cycle gets wrong about mid-market businesses
Mid-market leaders don't need a foundation model strategy. They need a first win. The gap between the two explains most failed AI projects we see.
ExploreA practical automation maturity model for mid-market operations
Most automation maturity frameworks are written for enterprises. This one is built for the reality of mid-market operations teams.
ExploreWe sunset Casetrack. Here is what 18 months of running production legal AI taught us.
We sunset Casetrack in April 2024. The honest reason is that the services work was earning its keep faster than the product was. Here is what 18 months of operating it taught us, and what it changed about how we work with mid-market law firms now.
Explore
How to launch without a product
Testing a business hypothesis and delivering core value to users without building an app. Why the no-product launch beats a traditional MVP in almost every process-centric business.
Explore