Connectors are the hard part of retrieval

If you lead engineers, you already know the temptation: solve the hard part yourself. That instinct fights connectors are the hard part of retrieval.

Build evals early. Without them you are arguing about vibes while users lose trust.

A RAG system without an owner for freshness is a demo with a long half-life of embarrassment.

Process should be light enough to change. If your AI workflow cannot be updated when a model, connector, or compliance rule changes, you do not have a workflow — you have a ritual.

Retrieval quality is an operations problem: stale sources, missing permissions, and chunking that shreds meaning.

Connectors and access control are where RAG projects die. Treat them as product work, not glue.

When agents join the loop, treat them like junior systems: limited privileges, explicit tools, budgets, and a human who owns the outcome. Autonomy without audit is just distributed risk.

On connectors are the hard part of retrieval, the leadership move is to make the invisible visible: ownership, verification, and the path for the next person.

Sustainability shows up as fewer retries, right-sized environments, and CI that does not burn cycles for vanity. Efficiency is operational maturity.

I prefer written decisions over verbal ones. Memory is a poor archive, and AI tools make fluent improvisation cheap — which raises the value of durable context.

None of this requires a new framework brand. It requires attention, a short feedback loop, and the humility to change process when agents join the workflow.

Lead for continuity. Leave systems and people that still work when you are not in the room.