Eddie's Voice Test™ (EVT™ )

 

When evaluating non-native English speaking (NNES) agents who serve native English speaking (NES) customers, your CSAT and NPS metrics rely heavily on effortless, frictionless communication. Eddie’s Voice Test™ (EVT™) by BK One Learning is a highly targeted diagnostic tool designed to measure and map the exact communication gaps between your agents and your customers.

​

The Core of EVT™

 

EVT™ functions as a structured Accent Neutralization Evaluation. It systematically breaks down an agent's speech to assess how they articulate specific segmental phonemes. To ensure comprehensive analysis, the test evaluates the physical production of sound across all articulatory areas, ranging from bilabial sounds (utilizing both lips) all the way to glottal sounds (originating between the vocal folds).

​

Intelligibility and the 5-Point Scale

 

Customer effort—such as forcing a caller to ask an agent to repeat themselves—severely damages CSAT. To address this directly, EVT™ scores pronunciation strictly based on intelligibility:

• Score 5: The agent exhibits no errors in segmental phoneme pronunciation, achieving a Native-English level.

• Score 4: There is a slight problem with pronunciation, but the agent's speech remains fully intelligible.

• Score 3: The agent demonstrates moderate difficulty, remaining intelligible but requiring one repetition.

• Score 2: There is a significant problem with pronunciation, and the speech is only intelligible after more than two repetitions.

• Score 1: The agent cannot pronounce the phoneme at all and is not intelligible.

 

Comprehensive Phonetic Breakdown

 

Rather than generic listening tests, EVT™ categorizes exactly which sound types need correction:

• Consonants: The tool evaluates Stop-Plosive, Fricative, Affricate, Nasal, and Glide sounds.

• Vowels: It formally assesses Front, Central, Back, Higher Back, and Diphthong sound production.

 

Practical Error Identification

 

The evaluation maps common NNES pitfalls against specific example words, tracking exact points of failure. For example, trainers can document if an agent drops the ending "t" (causing "what" to sound like "baa"), or if they pronounce the initial "v" in "victory" like a "w". By pinpointing whether an agent shrills the "r" in "prefer" or turns the short "i" in "him" into "heem", your training teams can deploy highly precise, data-driven neutralization strategies to improve customer interactions.