Level 3: Emerging Structure

Score range: 75–99

At Level 3, your team has started building consistent practices. There are shared approaches, some documentation, some review. AI is genuinely useful — but it’s not yet reliable. Quality is inconsistent, and the team knows it.

What This Level Looks Like

Level 3 is where most professional teams land after a year or two of intentional AI adoption. You’ve moved past individual experimentation. You have some shared vocabulary, maybe a prompt library, maybe a review step. But the practices aren’t fully embedded — they depend on remembering to use them, not on systems that make them inevitable.

AI output typically requires moderate revision before it’s usable. The team could tell you what makes a good AI output; it couldn’t always tell you why a specific output missed.

Your Core Risk

The primary risk at Level 3 is practice decay. The disciplines your team has built are real, but fragile. Under deadline pressure, they get skipped. When a key person is absent, they don’t happen. The team is one sprint away from sliding back toward Level 2 behavior.

A secondary risk: governance is likely underdeveloped. Your team is using AI more confidently — but without clear guidelines on what AI should and shouldn’t decide, the confidence may be outrunning the guardrails.

Your Core Strength

You’ve done the hard part — building practices from scratch. The next phase is about making those practices durable: encoding them in systems, tools, and team agreements rather than individual memory and goodwill.

Your Next Move

Pick one AI practice your team does inconsistently and make it structural. Put it in a checklist. Add it to your definition of done. Schedule a monthly review of AI output quality. Structural beats behavioral every time — you’re moving from “we try to do this” to “our process requires this.”

One thing: Identify which of your five dimension scores is lowest. That’s your highest-leverage investment right now.