What comes back

This is what you get back

A candidate speaks and writes for about twenty minutes. A few minutes later, this lands on your dashboard: a CEFR level, the ten scores behind it, feedback a hiring manager can act on, and every answer the candidate gave. Here is one, in full.

Free trial, ten assessments, no card. The report you get is this one with your name on it.

Sample resultNot a real candidate

Amara Okafor

ScoredSpeaking & Writing

amara.okafor@example.com · Customer Support Specialist · REQ-2041

B2

Upper Intermediate

64/100 overall

Language
English
Duration
21 min
Completed
14 October 2026
Tasks
6 of 6
Scored in
4 min

What B2 means

  1. A1
  2. A2
  3. B1
  4. B2
  5. C1
  6. C2

Can interact with a degree of fluency and spontaneity that makes regular interaction with native speakers possible.

Speaking

4 of 4 tasks

63B2
Pronunciation62
Fluency58
Grammar66
Vocabulary61
Coherence64
What went well
  • Held the thread of an argument across the 80-second answer and recovered naturally after losing a word.
  • Vocabulary suited a professional register throughout, including the situational task.
Opportunities
  • Hesitation before complex clauses slowed the pace in the extended response.
  • Stress fell on the wrong syllable in a few longer words, without obscuring the meaning.

Writing

2 of 2 tasks

65B2
Task response68
Organisation64
Vocabulary60
Grammar63
Conventions71
What went well
  • Covered every point the task asked for, in a clear order, with a register that suited a message to a manager.
  • Paragraphing made the longer piece easy to follow.
Opportunities
  • Article use was inconsistent ("the customer" and "customer" within one paragraph).
  • Leaned on the same two linking words; a wider range would read as more fluent.

Integrity

Nothing to call out

Ordered with integrity checks on. Pasting was blocked in written answers; everything else is recorded and shown here as observations, never as a verdict.

Times switched away
1
Time away
12s
Pastes blocked
0
Keystrokes
2,140
Corrections
96
Question copied
0

A second display was connected. Reported as a fact, not flagged: a docked laptop is the normal office setup.

The answers

Speaking · Extended response1:14

“Describe a time you had to explain a complicated process to someone who was frustrated.”

Transcript “So, last year a customer called about a refund that was, um, stuck between two systems. I could see she was already frustrated, so the first thing I did was tell her exactly what I was going to check, step by step, before I checked it…”

Writing · Professional

“Write to your manager proposing one change to how your team handles urgent requests.”

“Hi Daniel, I would like to propose a small change to how we handle urgent requests. At the moment the customer has to explain their problem twice, once to the first agent and again when it is escalated…” 214 words

Reading and listening results sit on the same page when ordered.Visible to your team only. Candidates never see their scores.
  1. One level, one score

    The headline is a CEFR level, A1 to C2, with its name and the 0 to 100 score behind it. B2 means the same thing for every candidate, in every language, on every skill, which is what lets you write "B2 or above" in a job description and hold every applicant to it.

  2. What the level means

    The level is placed on the six-step scale and described in the Council of Europe’s own terms, so a hiring manager who has never heard of the CEFR still knows what the candidate can do.

  3. A card per skill, five dimensions each

    Speaking is scored on pronunciation, fluency, grammar, vocabulary and coherence; writing on task response, organisation, vocabulary, grammar and conventions. A section score cannot sit far above its weakest core dimension, so a wide vocabulary never papers over grammar a listener would struggle with.

  4. Observations, never a verdict

    When integrity checks are on, the report shows what was recorded: time away from the tab, blocked pastes, keystrokes and corrections. There is deliberately no integrity score. A tab switch is a notification far more often than it is cheating, and a number would be read as proof.

  5. The answers themselves

    Every recording and every written response is on the report with the question it answered. The AI’s feedback is a summary; the evidence is one click away for anyone who wants to hear the candidate for themselves.

The ten scores behind the level

Each skill is scored on five dimensions, 0 to 100. The level is derived from them, so when two candidates land on the same level the dimensions are what tell them apart.

How the scores become a level
Speaking
Pronunciation
Clarity of sounds, stress and intonation. Accent is not penalised.
Fluency
Natural flow and pace, and how much hesitation interrupts it.
Grammar
Sentence structure, tense and accuracy in speech.
Vocabulary
Range and precision of word choice, and whether it suits the context.
Coherence
Relevance to the question, and how the ideas are organised.
Writing
Task response
Whether the response addresses every part of the prompt.
Organisation
Logical structure, paragraphing and flow.
Vocabulary
Range and precision of word choice, and whether it suits the register.
Grammar
Accuracy, sentence variety and complexity.
Conventions
Spelling, punctuation and formatting.

What the report will not do

A result that people make hiring decisions on has to be careful about what it claims. Four things this report leaves out on purpose.

It will not invent a level for a non-attempt
An answer in the wrong language, or an answer to a different question, gets "Not assessable" and the reason, not an A1. A1 is a real level with a real descriptor, and a candidate who did not attempt the task has not earned it.
It will not give you a pass or fail
You decide what level a role needs; the report tells you what level the candidate has. The two are kept apart so that the same result can serve a role that needs B1 and one that needs C1.
It will not penalise an accent
Pronunciation is scored on whether the candidate is understood: clarity, stress and intonation. Where they learned the language is not a dimension.
It will not show the candidate their scores
The candidate sees a thank-you screen. The report is for the people making the decision, and what they share with the candidate is their call.

Where it goes

The page you just read is the dashboard view. The same levels and scores reach your applicant tracking system the moment scoring finishes, by webhook, so the result can sit on the candidate's record without anyone copying it across.

  • Dashboard: every assessment, filterable by status, level and language.
  • Webhook: one signed POST when the result is ready, shown here.
  • API: the full breakdown, per dimension and per part, on request.
  • Email: your team is told when a candidate finishes.
API documentation
POST https://your-ats.example/evalingoassessment.scored
{
  "event": "assessment.scored",
  "data": {
    "assessmentId": "cm2x9k4p70001",
    "externalRef": "REQ-2041",
    "candidate": { "email": "amara.okafor@example.com", "firstName": "Amara" },
    "result": {
      "overallLevel": "B2",
      "overallScore": 64,
      "speaking": { "level": "B2", "score": 63 },
      "writing":  { "level": "B2", "score": 65 }
    },
    "integrity": { "concerns": 0, "notices": 0 },
    "detailsUrl": "https://evalingo.ai/api/v1/assessments/cm2x9k4p70001"
  }
}

Questions people ask

Is this a real result?
No. Amara Okafor is invented, and so is every answer on the page. The layout, the dimensions, the feedback format and the integrity section are exactly what a real result contains; the numbers are ones the scorer produces every day, chosen to show a typical B2.
How long after the candidate finishes does this arrive?
Speaking and writing are scored by AI as each answer is submitted, so the report is usually complete within a few minutes of the last answer. Reading and listening are calculated the instant the candidate finishes.
What does a reading or listening result look like?
A level and a score on the same scale, how many questions were answered correctly in each part, and which second-part route the candidate’s first answers placed them on. Every screen the candidate saw is reviewable with the answer they chose and the correct one. There is no AI feedback, because nothing in those tests is judged by a model.
Can I share the report with a hiring manager?
Yes. Anyone on your Evalingo team can open it, and the API and webhook carry the same levels and scores into your applicant tracking system, so it can live wherever the rest of the candidate’s record does.
Can I see one with my own answers?
That is what the free trial is for. Sign up, send the assessment to yourself, and the report you get back is this page with your name on it. Ten assessments, no card.

See it with your own answers

Sign up, send the assessment to yourself, and read your report in about twenty-five minutes. Then send it to a candidate.

No credit card. No sales call.