AI doesn't fail because of prompts. It fails because of the environment behind them.
When AI encounters a low-quality codebase or an unsuitable SDLC, it amplifies existing engineering problems instead of solving them. Discover whether your repository is helping AI—or holding it back.
Most teams have the tools. Few have built the environment those tools need to work inside.
Supervising
AI capability is built into how the organisation works. New engineers inherit it, the system improves itself.
Orchestrating
The same intent, written down once, drives consistent work across agents - and across time.
Regulating
The environment carries the rules. Conventions and guardrails live in the repo, working whether you're watching or not.
Commanding
You instruct more precisely and check what comes back. Tests, reviews, and judgment catch the mistakes before they ship.
Dictating
Every task starts fresh. You type the instructions, you carry the context, and the results swing with how well you explained it.
The assessment places your team on this scale - then shows you exactly what to build to reach the next level.
A scan, a score, and a plan - in minutes.
The assessment runs directly on your repository. No forms, no manual review. It reads what's there, scores what it finds, and tells you what to do next.
Point it at your repository
Install the AI Literacy Superpowers skill via Claude Code or GitHub Copilot. Run the assessment command against any software repository - existing projects included.
The scan runs automatically
The assessment reads your codebase across five dimensions: context files, conventions, architectural guidance, guardrails, and agent readiness. It maps what you have against what each level requires. Everything runs locally - no source code, no scan results, and no repository data ever leaves your machine.
You receive your score and plan
Your team is placed on the five-level scale – from Dictating to Supervising. You get a prioritised improvement plan: a ranked list of what to address first to reach the next level and unlock real AI productivity gains. See a sample report.
Five levels of AI readiness - where does your team operate?
Level 1 = Dictating · Level 5 = Supervising
1
Dictating
Every task starts fresh. You type the instructions, you carry the context, and the results swing with how well you explained it.
2
Commanding
You instruct more precisely and check what comes back. Tests, reviews, and judgment catch the mistakes before they ship.
3
Regulating
The environment carries the rules. Conventions and guardrails live in the repo, working whether you're watching or not.
4
Orchestrating
Specifications do the coordinating. The same intent, written down once, drives consistent work across agents - and across time.
5
Supervising
AI capability is built into how the organisation works. New engineers inherit it, the system improves itself, and your job is shaping how it scales.
Find out where your codebase stands
Created by Russ Miles.
Built on real-world engineering consulting by TechTalk.
For over 30 years we have been the sparring partner Austrian and DACH organisations turn to when technology waves reshape their industry – embedding with leadership and engineering teams, teaching them to navigate the change, and helping turn disruption into lasting competitive advantage.
We have navigated every technology wave – PC & Office, Web & Cloud, and now Agentic AI – and the organisations that move with a partner who has crossed them all before are the ones that come out ahead.
Have a question?
Whether you want to understand what the results mean, or talk through what implementation actually looks like — just reach out. We're happy to help.
Two outputs.
All from one scan.
The assessment doesn't just tell you where you are - it shows you exactly what to do next.
Your AI Readiness Score
A clear placement on the five-level scale - from Dictating to Supervising - with a breakdown across each dimension: context, conventions, architectural guidance, guardrails, and agent readiness.
A Prioritised Improvement Plan
A ranked list of what to build first - ordered by what your team needs to develop and what your organisation needs to provide to reach the next level.
Need help turning the plan into action?
We embed with engineering teams to implement the recommendations, close the gaps, and build the practices that make AI a genuine multiplier - not a debt accelerator.
Find out where your codebase stands
Run the assessment in minutes. Find out whether your team is Dictating, Commanding, Regulating, Orchestrating - or Supervising. Free, with no signup required.
