The Emeri Standard
How the grade is made.
Emeri measures how well you work with AI, against a published standard that is the same for everyone. This page explains exactly what we test, how the bands are defined, how grading works, and — just as important — what a grade does not claim.
The six competencies
Every assessment scores the same six competencies. Together they describe what it means to be genuinely good with AI at work — independent of any particular tool or model.
- Asking well
- The quality of your instructions to AI. Framing a task clearly, giving structure and constraints, and asking again when the first answer misses — so the model works on the problem you actually have.
- Checking the work
- Judgement about output. Reading a response critically, catching errors and confident nonsense, verifying claims, and knowing which parts to trust and which to check by hand.
- Putting it together
- Working across tools and steps. Breaking a job into stages, moving between AI and everything else in your workflow, and stitching the pieces into a result that holds together.
- Getting real work done
- Finishing real work. Taking a task from prompt to something a colleague could use — not an impressive draft, but the done thing, in the form the job actually needs.
- Giving good context
- Supplying the right inputs. Giving AI the information, data, examples, and constraints it needs to be useful, and recognising when it is missing something important.
- Keeping it safe
- Responsible use. Protecting private and proprietary information, respecting IP, and knowing where AI should not be trusted with a decision on its own.
The full framework, including behavioral markers for each band, is published in the AI Skills Assessment Framework.
The bands: E1 to E4
Your overall result is a band from E1 to E4. Bands describe behaviour — what someone at that level actually does with AI — not a raw score. The scale is deliberately narrow so each step means something.
E1 · Foundational
Can get useful results from AI on familiar tasks with clear instructions, but leans on it without always checking the work. Foundational, and a real starting point.
E2 · Proficient
Works with AI reliably across everyday tasks — asks well, catches obvious errors, and produces work that stands on its own. The proficient professional standard.
E3 · Distinguished
Uses AI with judgement and range: combines tools, handles ambiguity, verifies what matters, and delivers work that raises the bar. Distinguished.
E4 · Exceptional
Sets the standard others learn from. Fluent across every competency, deliberate about where AI helps and where it does not, and consistently excellent under pressure. Exceptional.
How grading works
The free check is a calibrated, multiple-choice assessment. It places you on the E1–E4 scale and gives you a skill profile across the six competencies in about fifteen minutes. It is a real signal, and it is free.
Certification is more demanding. It draws fresh questions you have not seen, and adds free-response scenarios — open-ended tasks graded against published rubric criteria. Those rubrics are written by people; an AI examiner applies them consistently to every submission. Your certified grade blends the two parts: 40% multiple choice, 60% scenarios, because judgement on open work is what most separates the bands.
Your result carries two reference points. Your percentile compares your score against everyone who has taken the same assessment, recalculated as the population grows — so it always reflects the current field rather than a fixed curve. Alongside it we publish an estimated standing across the broader professional workforce, clearly labelled as an estimate: exam takers are a self-selected, AI-experienced group, so your position among working professionals at large is meaningfully higher than your rank among them.
Every certified result also carries a short written note from the examiner — a specific account of what your work demonstrated, attached to your credential where a viewer can read it. It is generated from your actual answers, never from a template.
The pass bar
Every band from E1 to E4 is a certified grade — E1 · Foundational is an earned credential, not a consolation. But certification has a floor: a sitting that doesn't demonstrate even foundational competence is not graded at all. No badge, no band, no percentile — you get your examiner's written feedback and a pointer to exactly the skills to study first. One retake is included with every purchase for exactly this case, and our free study guides cover every topic on the exam.
Keeping it fair
Questions are drawn from larger pools and shuffled, so no two sittings are identical. A certified grade cannot be retaken on the same payment to fish for a higher band — the included retake exists only below the certification bar, and it draws fresh questions. These are ordinary integrity practices, and they are why an Emeri grade is worth trusting.
What Emeri does not measure
A grade is a strong signal, not a full picture of a person. Emeri does not measure your overall job performance, your domain expertise, or your value as a colleague. It does not test memorised trivia about specific model versions, which change monthly — though it does test the durable judgment of matching the right class of model to the job, spending capability where the stakes demand it and economy where they don't. And it cannot see how you will grow. It measures one thing well: how capably you work with AI, today, against a standard that is the same for everyone.
Where to start
The free check places you on this scale in about fifteen minutes — no account, and nobody sees your grade but you. Certification is the same standard, examined: one sitting, one payment, a grade anyone can verify.