A comprehensive technical breakdown of Trio's 7-part 63-question assessment architecture, acoustic neural diagnostics, NLP syntax tree parsing, configurable decision matrix, and CEFR-aligned composite weighting algorithms.
Trio evaluates English proficiency not as an arbitrary flat grade, but across four mathematical linguistic dimensions calibrated to international CEFR workplace standards.
Neural speech models process raw audio spectrograms at 16kHz to diagnose pronunciation clarity, rhythm, cadence, and vocal hesitation.
Measures auditory attention, phonological loop retention, and conversational parsing using strict single-play locked audio dialogues.
Evaluates syntactic hierarchy, subordinate clauses, subject-verb agreement, and sentence structure building under time constraints.
Measures the sophistication, contextual suitability, Type-Token Ratio (TTR), and pragmatic suitability of candidate expression.
Every candidate completes a standard 63-question assessment battery divided into 7 psychometrically calibrated parts, designed to systematically test every layer of spoken and written English.
Candidates are presented with 3 scrambled phrase groups (e.g. ["was delayed", "the morning dispatch", "by heavy traffic"]) and must tap the blocks in the exact sequence to construct a grammatically correct, meaningful English sentence.
Longest Common Subsequence (70%) + Levenshtein Word Distance (30%) with non-linear strict power curve.
Syntactic order, dependency parsing, phrase structure rules.
Exact match: 100 pts • Partial order: 50–90 pts • Incorrect: 0 pts.
Candidates evaluate a target vocabulary prompt and select the closest semantic match from 4 carefully calibrated options spanning business terminology, academic verbs, and workplace collocations.
Normalized exact choice match against CEFR word-frequency tiered key.
Contextual semantic precision, collocation awareness, advanced lexicon.
Correct option: 100 pts • Incorrect option: 0 pts.
Candidates read aloud full operational policy passages and customer interaction texts into their microphone. The AI engine captures acoustic telemetry, evaluates phone-level alignment, and measures speech tempo.
Acoustic Word Error Rate (WER $\le 0.15$), WPM cadence (120–160), and neural acoustic confidence multiplier.
Pronunciation clarity, connected speech rhythm, vowel formants, stress.
Continuously scored 0–100 based on $(1 - \text{WER} \times 2.2)^{1.5} \times \text{Confidence}$.
Candidates listen to short spoken sentences with strict single-play enforcement (Audio plays ONLY ONCE). Candidates must immediately type the exact sentence with correct spelling and punctuation.
String Levenshtein edit distance & Word Error Rate: $(1 - \text{WER} \times 2.0)^{1.6} \times 100$.
Phonological loop storage, auditory decoding, orthographic transcription.
Threshold $\ge 75$ for full credit; scaled continuously based on edit distance.
Candidates listen to audio recordings of workplace consultations, customer escalations, and dispatch announcements (1 playback only), then answer multiple-choice questions assessing both main ideas and granular details.
Objective key matching with single-play timestamp validation.
Gist inference, rapid information filtering, international accent tolerance.
Correct option: 100 pts • Incorrect option: 0 pts.
Candidates read technical operations memos, privacy regulations, customer dispute policies, and workflow procedures, answering contextual questions that test comprehension depth beyond simple keyword search.
Multi-option contextual verification engine.
Skimming, scanning, logical deduction, nuanced business policy interpretation.
Correct option: 100 pts • Incorrect option: 0 pts.
Candidates are given real-world customer service scenarios (e.g. "A customer expresses frustration regarding a delayed pickup due to unexpected rain. Write a polite 2-3 sentence email response...") and compose formal, polite responses.
Combined ratio of Length Volume (50%) + Type-Token Ratio (30%) + Domain Keyword Coverage (20%).
Professional register, diplomatic tone, syntactic complexity, lexical variety.
Graded 0–100 via $(\text{Length} \times 0.50 + \text{TTR} \times 0.30 + \text{Keywords} \times 0.20)^{1.6} \times 100$.
How raw item scores transform into CEFR certifications and automated hiring decisions via Trio's multi-tenant configurable decision matrix.
Every response across all 63 items is evaluated by specialized micro-engines (Speech WER, Levenshtein, MCQ verification, NLP TTR) to generate exact raw scores from 0 to 100.
Scores for each of the 7 sections ($S_A, S_B, \dots, S_G$) are averaged and combined into the 4 primary competency pillars (Speaking 35%, Listening 25%, Grammar 20%, Vocabulary 20%).
The weighted composite score is mapped against standardized psychometric scales (Trio Scaled Score 20–80, CEFR Tiers A1–C2, Global Proficiency Bands) and recruiter-tuned hiring recommendation cutoffs.
Test the composite formula in real time: Adjust section sliders to observe how Trio's weighting algorithm calculates composite scores, standardized equivalencies, and assigns hiring recommendations.
How Trio composite score ranges translate into Common European Framework of Reference (CEFR) levels and corporate hiring readiness.
| CEFR Level | Score Range (0–100) | Trio Scaled (20–80) | Global Band | Hiring Recommendation | Workplace Proficiency Profile |
|---|---|---|---|---|---|
| C2 | 95 – 100 | 78 – 80 | Band 8.5 | Strong Hire | Mastery — Flawless, native-like spontaneous expression, idiomatic precision, and executive nuance. |
| C1 | 85 – 94 | 68 – 77 | Band 8.0 | Strong Hire | Effective Operational Proficiency — Complex arguments, fluent discourse, high grammatical accuracy. |
| B2 | 70 – 84 | 54 – 67 | Band 6.5 – 7.0 | Hire | Vantage / Upper Intermediate — Independent workplace communication with native speakers without strain. |
| B1 | 55 – 69 | 42 – 53 | Band 5.0 – 6.0 | Conditional Hire | Threshold / Intermediate — Routine workplace tasks, basic client support, simple descriptive email drafting. |
| A2 | 40 – 54 | 30 – 41 | Band 4.0 | Do Not Hire | Waystage / Elementary — Basic routine interactions, slow speech comprehension, frequent grammatical breakdowns. |
| A1 | 0 – 39 | 20 – 29 | < Band 4.0 | Do Not Hire | Breakthrough / Beginner — Isolated memorized words, severe hesitation, unable to sustain workplace dialogue. |
