Each dimension is assessed independently on a five-stage developmental scale. Your profile across all six is far more useful for growth than a single aggregate score.
How clearly and specifically you describe goals, context, and constraints in your prompts. This is the single most important variable in output quality — more than the model, more than the platform.
"Am I giving the AI everything it needs — or expecting it to guess?"
How you respond to errors and unexpected outputs. Do you adapt, interrogate, and learn — or abandon and restart? Expert AI users treat every unsatisfying output as diagnostic data, not a dead end.
"When it doesn't work, do I understand why — or do I just try again?"
Your ability to actively steer the AI rather than passively follow where it leads. Are you the author of the interaction or a passenger? Direction is the difference between AI as a tool and AI as a crutch.
"Am I steering this conversation — or being steered by it?"
Your critical assessment of AI outputs. Can you identify inaccuracy, bias, or misread intent? Evaluation is where human expertise and AI capability meet — and where novice users are most vulnerable.
"Can I tell when it's wrong — especially when it sounds confident?"
Understanding what to outsource to AI versus what to own yourself. This is the workflow intelligence dimension — knowing which tasks benefit from AI assistance and which require human judgement, creativity, or accountability.
"Am I delegating the right things — or outsourcing my thinking?"
Your conceptual grasp of how AI systems work. Not coding — understanding. What is a context window? Why does prompt structure matter? What causes hallucinations? This knowledge underpins every other dimension.
"Do I understand why AI behaves the way it does — or does it feel like magic?"
Each AIDED-T dimension is rated independently on a scale drawn from the Dreyfus model of skill acquisition. You might be highly competent at Articulation while still at Novice level on Technical Literacy. That profile is far more useful for development than a single aggregate score — and it shows you exactly where to focus next.
Credential platforms like AI CRED weight Prompt Mastery at 40% of their total score — and they are right to. Prompt quality is the primary variable in AI output quality. AIDED-T agrees. But it disaggregates Prompt Mastery into two distinct dimensions (Articulation and Direction) because these are separate skills that develop at different rates and require different kinds of practice.
| What it measures | Credential score approach | AIDED-T approach |
|---|---|---|
| Primary purpose | Benchmark current performance for employers | Map development to guide growth |
| Prompt skill | Single Prompt Mastery score (40%) | Split into Articulation + Direction (independent) |
| Neurodivergent learners | Scaffolding penalised or invisible | Scaffolding recognised as a sophisticated adaptive strategy |
| Output | Single number between 1–10 | Six-dimension profile — 5 stages each |
| Use case | Signal to external audience | Personal and organisational development roadmap |
For neurotypical learners, learning velocity is an interesting metric. For ADHD learners, it is arguably the most diagnostically significant variable of all. Working memory limitations mean complex tasks require more external scaffolding to sustain. Inhibitory control differences mean starting is often harder than continuing. AIDED-T names and tracks these patterns — rather than obscuring them behind an aggregate score. Knowing how to use structure, checklists, and step-by-step prompts is itself a sophisticated skill. A good developmental framework should recognise that, not penalise it.
A new wave of AI credentialling platforms has arrived. You can now take a live conversational assessment, receive a numerical score, and add it to your LinkedIn profile. The ambition is serious and the need is real. AI fluency is becoming one of the most valuable professional skills of the decade.
But there is a problem with the credentialling approach — and it matters most for the people who stand to benefit most from working with AI.
Professional benchmarks are designed to answer a single question: how good are you right now? That is a useful question if you are an employer screening candidates. It is much less useful if you are trying to understand how to get better.
More critically, a snapshot score treats AI fluency as a fixed skill — like typing speed. But working productively with AI is a dynamic, evolving, deeply personal capability. It is shaped by your domain knowledge, your communication habits, your tolerance for ambiguity, and — crucially — by the cognitive profile you bring to the interaction.
For neurodivergent professionals in particular, a credential-focused assessment can actively obscure what is most important about how they learn.
Rather than asking an AI to simply generate a framework, I asked Claude to go into research mode first — to consult academic literature across pedagogy, educational psychology, computer science education, and human-AI interaction. To build something evidence-based before we discussed it.
The synthesis drew on Bloom's Digital Taxonomy, Kolb's Experiential Learning Cycle, the Dreyfus model of skill acquisition, and emerging research on ADHD and AI-assisted executive function scaffolding. It connected these to real commercial frameworks already in circulation, then identified the gaps.
The acronym itself was a byproduct of the process: Claude identified five core dimensions, noticed the letters spelled AIDED, recognised this was apt for an AI-learning framework, and built the sixth dimension — Technical Literacy — to complete it. That is what genuine human-AI intellectual collaboration looks like.
The complete published version includes the Momentum Dimension section — why ADHD changes everything about how we measure AI learning velocity.