How do you make knowledge more valuable across an organization? At TextLayer, we're exploring how sharing expertise, challenging assumptions, and learning from each other can improve technical decisions and help good ideas get better.

Can a small AI model navigate an IBM mainframe? Gareth Sharpe experiments with TypeSafe AI's Jev and Claude Haiku to explore a different approach to process automation, replacing fixed navigation scripts with model-driven decisions.

TextLayer tested TypeSafe AI’s Jev across classification, routing, validation, ranking and prediction to understand where fast, low-cost decision models fit alongside code, LLMs and human judgment.

Your first AI deployment should reveal what customers will actually use. Accelerate the production lessons others can teach you, and focus on the customer learning that should shape what you build next.