AI in the editor,humans in the loop.
Most engineering orgs say they’re AI-native. The instrument disagrees. Drag the marker below to feel the difference between manual, assisted, augmented, and autonomous — then take the diagnostic to find out where your team really sits.
Manual.
The ceremony of pretending AI hasn’t happened. Pull requests open, sit, get reviewed by humans. Estimates land in spreadsheets. Pasting into a chatbot is a personal productivity habit, not an organisational practice. Cycle time is whatever it always was.
Assisted.
The keystroke gets help. The process does not. AI sits inside the IDE; it does not sit inside the workflow. Code is generated and then carefully human-reviewed; specs are still hand-typed. Adoption is per-engineer, not per-team.
Augmented.
AI is an organ, not an accessory. Specs are drafted by it, code is reviewed by it, runbooks are queried by it. Humans set direction; the system handles the typing, the pattern-matching, and the first cut. Cycle time visibly shortens. People notice.
AI-Native.
Autonomous loops own routine work end-to-end. Self-healing systems. Humans set objectives, audit results, own the policy. The org operates as a director, not a do-er. You can feel it on a Friday afternoon.