Talk to an engineerIntro call
Start the assessment

How ready is your team for AI in production?

Ten questions, about four minutes, no email address. Answer honestly rather than aspirationally - the result is only useful if it describes the team you have. Questions are grouped into the four areas we work in, so the score tells you not just how ready you are but where the gap sits.

In short

This free self-assessment from 23Minds helps a team judge how ready it is to run AI features in production. It contains ten questions across four areas: governance and adoption, evaluation and security, building and integration, and operations and measurement. Each answer scores zero to three points, giving a maximum of thirty, and the result places the team in one of four readiness bands with specific next steps. The tool is self-service and free, it collects no personal data and stores no answers, and every result links to the practices most relevant to that band.

Train, Adopt and Govern

Do your teams have a written policy for using AI tools at work?
How do people in different roles learn to work with AI tools?
Do you know which AI systems and AI features are in use across the organisation?

Evaluate, Audit and Secure

How is AI-generated code handled before it is merged?
Have your LLM-facing applications been tested for prompt injection and data leakage?
What has to be true before an AI feature is allowed into production?

Build, Automate and Modernize

How far have you gone with agents that call tools and act on your systems?
How are AI features connected to the systems that actually run your business?

Operate, Measure and Maintain

What can you see about what your AI features do in production?
How do you control what AI features cost you?
FAQ

Questions teams ask.

No. The assessment runs in your browser, we store no answers and there is no form to fill in. If you want to discuss the result, you contact us - not the other way round.
Someone with a view across both delivery and operations: an engineering manager, a CTO, a head of delivery or a product owner close to the technical work. If two people from the same organisation disagree on an answer, that disagreement is itself a finding worth discussing.
Not on its own. A low score matters when the exposure is high - regulated data, money movement, decisions affecting people. A startup validating an idea and a public sector body processing citizen requests should read the same score completely differently.
Accurate enough to tell you where to look, not accurate enough to plan a budget from. It is a triage instrument: it points at the weakest of the four areas and gives you a vocabulary for the internal conversation. A real assessment reads your code, your pipeline and your architecture.
Talk to an engineer and bring the result. We use it as the starting point for scoping, which usually turns into either a fixed-scope assessment with a written report, or a training and policy engagement, depending on which area scored lowest.

Talk to an engineer.

Describe where you are with AI. A reply within one business day.

Talk to an engineer
Szczecin - ul. Wawrzyniaka 6WWarszawaZielona GóraKraków