Skip to main content

​
Overview

Custom pronunciations let you control how the AI interviewer says specific words out loud. You define a rule with two parts: the term as it is spelled, and how it should be spoken. Rules apply organization-wide to AI interview calls, so every job uses the same pronunciations.
Only the spoken audio changes. Transcripts, notes, summaries, and anything else written keep the original spelling.

​
When to use it

Add a rule whenever the AI mispronounces a word that matters to your candidates:
  • Company and brand names — your organization, your client names, product names
  • Acronyms — terms that should be read letter by letter, or read as a word
  • Uncommon terms — industry jargon, certifications, or location names the AI gets wrong
If a term already sounds correct, leave it alone.

​
Where to configure

Navigate to Admin > Interview Settings > Pronunciations.

​
Adding a rule

  1. Go to Admin > Interview Settings > Pronunciations.
  2. Enter the term under spelled as — write it exactly as it appears in your questions and scripts.
  3. Enter the respelling under spoken as.
  4. Select hear it to listen to a sample of the respelling.
  5. Adjust and preview again until it sounds right, then save.
By default a rule matches any capitalization, so a rule for Nexora also matches nexora and NEXORA. Turn on Match case for a term whose casing carries meaning (for example an all-caps acronym), and it will only match that exact capitalization. Either way, matching applies to whole words and does not fire inside a longer word. Each preview plays two clips: one saying “This used to be pronounced as…” with the word spoken without your rule applied, then one saying “Now this will be pronounced as…” with the word spoken with your rule applied, so you can tell them apart by ear. Use the preview voice picker above the table to choose which voice reads the samples — it defaults to your organization’s default voice and only affects previews, not any job’s actual voice setting.

​
Writing a good respelling

The spoken as field is a plain-English respelling — spell the term the way it sounds, using ordinary letters. Phonetic notation is not supported. Practical tips:
  • Spell it how it sounds. Write what you would write if you were helping a colleague read it aloud for the first time.
  • Break it into syllables. Separate the parts with spaces so each one lands clearly.
  • Use spaces for letter-by-letter. To have an acronym read out one letter at a time, separate the letters with spaces.
  • Preview before you save. The respelling that reads best on the page is not always the one that sounds best, so use hear it to check.

​
Examples