12 Signals Your Team Is Performing AI Adoption, Not Aligning To It

A diagnostic for engineering managers who can’t tell whether their team is bought in - or just quiet.

By 2026, most engineering organizations have crossed some adoption threshold with AI coding tools. The dashboards are green. The visible resistance that defined 2023-2024 has largely faded. Most leadership teams declare victory and stop paying attention.

This is exactly the moment the dangerous problems start. The engineers who would have resisted hardest go silent. Teams that ship faster on paper accumulate a new, uncatalogued kind of debt. The senior engineers whose judgment held the system together quietly update their LinkedIn. The metrics keep looking fine through the exit.

The checklist flips the frame. Each signal is something you can observe on your team - not something your team tells you. Tick a box only if you can point to a recent, concrete example. If you’d have to guess, leave it empty. Every tick is a place the team may be complying instead of aligning, and a prompt for a question you probably haven’t asked yet.

What’s in the PDF

Twelve signals across six dimensions:

  1. Measurement & Visibility - adoption rate vs. impact rate; felt productivity vs. measured productivity.
  2. Quality & Technical Health - throughput without inspection; comprehension debt in AI-generated code.
  3. Team Dynamics & Culture - loud juniors / quiet seniors; PR review time drifting up after the tooling landed.
  4. Leadership & Accountability - who owns AI-generated code when it breaks; the AI champion with no authority.
  5. Organizational Structure - thinner management layers with no redefined mandate; the same two or three faces presenting every win.
  6. Expectation Management - projections that outrun measured results; honest conversations that only happen after an incident or an outsider.

Each signal gives you:

  • What you’ll see - the observable, not the self-report.
  • What it means - the second-order read.
  • Evidence - a cited data point (Jellyfish, Stack Overflow, LinearB, Gartner, Korn Ferry).
  • First move if it’s true - one concrete action to run this week.
  • A tick box and a notes line for what you actually observed.

How to score it

Count the ticks. The final page gives you a reading:

  • 0-3 - Mostly aligned. Check the ticked signals for blind spots anyway.
  • 4-6 - Mixed. Start with the lowest-numbered signal you ticked; measurement problems feed the rest.
  • 7-9 - Performing. You’re reading a dashboard your team has learned to keep green. Expect the cost to show up as quiet attrition before it shows up in metrics.
  • 10-12 - Deep performance. Reporting and reality have separated. The most useful next step is one honest conversation, not another tool or training.

The PDF is a fillable form in Adobe Acrobat or Reader (the score updates as you tick). In a browser preview or on a phone, count the ticks yourself.

What this isn’t

This doesn’t measure how much AI tooling you have, which vendors you use, or how many people finished training. It asks whether your management layer can see what’s really happening on your team and act on it. The teams that get lasting value from AI won’t be the ones that adopted fastest - they’ll be the ones whose managers could see clearly, measure honestly, and say what they found.

Get the diagnostic

Download the fillable PDF (8 pages, ~25 minutes of honest reflection).

Prefer to talk it through? Email anton@otomato.io with your count in the subject line.

Anton Weiss is a former CTO and longtime DevOps transformation consultant who coaches engineering leaders through cultural inflection points. More about Anton · Working with Anton.