At Databricks they’re responsible for the creation, care, and feeding of the AI agents we have in production. As these start to perform real use cases, you have to ensure they’re actually delivering the outcomes they were designed to do… and don’t drift. More often than not they’re resolving upstream data/context issues and now cost optimization with model selection and testing.
2027 will be the year of model caps & optimization. Local >> cheap >> premium, depending on the task, your role, and expectations. Tokenomics are coming.
“Prove to me you drove $100k with of value with your usage.”