AI

Can an agent tell you whether this project made money? Further than a year ago, not as far as the demo suggests. What sits in the gap, and who has to decide it.

Pure build is rarely practical for a midmarket firm; pure buy is rarely sufficient. The practical path is sequential: assess first, buy what is commoditized, build only what is proprietary.

The AI projects that stall don’t stall at the model. They stall at the business underneath it. Ingo Roemer on the readiness work that decides whether an AI project holds up: data, process, connection across systems, and the governance question most agent deployments ask too late.