Work with dictionary management

• Click Menu > Conversation Intelligence > Speech and Text Analytics > Dictionary Management.
• From the list that appears, search for a specific term, search for terms associated with a specific dialect, edit a term, and delete a term.

1. Click Menu > Conversation Intelligence > Speech and Text Analytics > Dictionary Management.
2. Click Add Term.
3. Select the transcription engine that you want to target:
   • Genesys Native Voice Transcription
   • Extended Voice Transcription Services (EVTS)
4. Add a new term for a Genesys native voice transcription dialect: 
   
   1. Under Term, add the term. The term cannot be more than 30 characters.
   2. Select a dialect.
   3. (Optional) In the Sounds Like field, type a word that sounds like the term to add and then press Enter. You can add up to 10 words. 
   4. If the Example Phrases field appears for the selected dialect, click Add Phrase and enter an example phrase that contains the dictionary term.
      Note: The Example Phrases field appears only for dialects whose transcription model uses example phrases. You cannot add more than 20 example phrases. The dictionary management system does not support special characters or capitalization. All entries are processed in lowercase and must contain only standard alphanumeric characters (a—z, 0–9). For English, Spanish, and French language variants that use Genesys Native Transcription, example phrases are not used and the field does not appear.
   5. (Optional) In the Display As field, enter exactly how you want the recognized term to appear in interaction transcripts.
      Notes: 

      • Use this field to specify acronyms, preferred capitalization, product or brand-specific spelling, numbers, punctuation, or other formatting.
      • You can enter characters from any character set in this field.
      • For example, enter ceo as the dictionary term and C.E.O. in Display As to display recognized instances of the term as C.E.O.
      • The Display As value can contain up to 50 characters.
5. In the Boost Value field, select the recognition weight for the dictionary term. A higher value increases the likelihood that the transcription engine favors the dictionary term when multiple interpretations are possible. The supported values are 1.0 through 10.0, in increments of 1. The default value is 2.0.
6. Click Add to Dictionary. For more information, see Dictionary management best practices.

1. Click Menu > Conversation Intelligence > Speech and Text Analytics > Dictionary Management.
2. From the list that appears, click the term you want to edit.
3. Make the desired updates. Depending on the selected transcription engine and dialect, you can update Sounds Like, Example Phrases, Display As, and Boost Value. Only fields supported by the selected transcription model appear.
4. Click Add to Dictionary.

1. Click Menu > Conversation Intelligence > Speech and Text Analytics > Dictionary Management.
2. From the list that appears, in the Action column, click Delete .

Dictionary entry status shows at an individual term level, whether it has been successfully applied, and can be expected to appear in conversation transcripts. Possible term statuses include:

• Pending: Waiting to be applied
• Active: Available in transcription
• Failed: Not applied due to error
• Invalid: Not supported or incorrectly formatted

Rules for EVTS dictionary entries

• No digits in terms (spell out numbers)
• No hyphens or unsupported special characters
• Acronyms must include periods (for example, F.B.I.)
• Enter multi-word terms with spaces (the system converts internally)

Character limits

• Term: 30 characters
• Display As: 50 characters

Validation and errors

The system validates dictionary entries when you add or edit them. An error can occur under these circumstances:

• The term includes unsupported characters
• The format is invalid (for example, incorrect acronym format)
• The dialect or engine combination is not supported

Validation applies to both the user interface and APIs.

• Dictionary entries are engine-specific
• Entries are not automatically migrated between engines
• Recreate terms when switching engines
• The same term can exist in both engines

Each supported language has a defined character set that the Native transcription engine recognizes. When you create Native dictionary terms, use only the characters allowed for your dialect. Using unsupported characters cause validation to fail.

For the Native transcription engine, character-set validation applies to:

• Term values
• Sounds Like values
• Example Phrases, for dialects where the Example Phrases field is available

Languages

English character set (en-US, en-GB, en-AU, en-ZA, en-IN)
Spanish character set (es-US, es-ES)
German (Germany) character set (de-DE)
German (Switzerland) character set (de-CH)
French character set (fr-FR, fr-CA)
Portuguese character set (pt-BR, pt-PT)
Italian character set (it-IT)
Dutch character set (nl-NL)
Polish character set (pl-PL)
Arabic character set (ar-001, ar-SA, ar-AE, ar-BH, ar-EG, ar-IL, ar-TN)
Hebrew character set (he-IL)
Korean character set (ko-KR)
Japanese character set (ja-JP)

English character set (en-US, en-GB, en-AU, en-ZA, en-IN)

For English dictionary terms, you can use the following characters in the Term and Sounds Like fields:

• Letters A–Z and a–z only
• Space (” “)
• ‘ (apostrophe)

Spanish character set (es-US, es–ES)

For Spanish dictionary terms, you can use the following characters in the Term and Sounds Like fields:

• A-Z
• a-z
• Space (” “)

You can also use the following Unicode characters in these fields:

For German (Germany) dictionary terms, you can use the following characters in the Term value, Sounds Like and Example Phrases fields:

• A-Z
• a-z
• Space (” “)
• ‘ (apostrophe)

For German (Switzerland) dictionary terms, you can use the following characters in the Term value, Sounds Like and Example Phrases fields:

• A-Z
• a-z
• Space (” “)
• ‘ (apostrophe)

French character set (fr–FR, fr-CA)

For French dictionary terms, you can use the following characters in the Term and Sounds Like fields:

• A-Z
• a-z
• Space (” “)
• ‘ (apostrophe)

You can also use the following Unicode characters in these fields:

Portuguese character set (pt-BR, pt-PT)

For Portuguese dictionary terms, you can use the following characters in the Term value, Sounds Like and Example Phrases fields:

• A-Z
• a-z
• Space (” “)
• ‘ (apostrophe)

You can also use the following Unicode characters in these fields:

Italian character set (it-IT)

For Italian dictionary terms, you can use the following characters in the Term value, Sounds Like and Example Phrases fields:

• A-Z
• a-z
• Space (” “)
• ‘ (apostrophe)

You can also use the following Unicode characters in these fields:

Dutch character set (nl–NL)

For Dutch dictionary terms, you can use the following characters in the Term value, Sounds Like and Example Phrases fields:

• A-Z
• a-z
• Space (” “)
• ‘ (apostrophe)

You can also use the following Unicode characters in these fields:

Polish character set (pl–PL)

For Polish dictionary terms, you can use the following characters in the Term value, Sounds Like and Example Phrases fields:

• A-Z
• a-z
• Space (” “)
• ‘ (apostrophe)

You can also use the following Unicode characters in these fields:

Arabic character set (ar-001, ar-SA, ar-AE, ar-BH, ar-EG, ar-IL, ar-TN)

For Arabic dictionary terms, you can use the following characters in the Term value, Sounds Like and Example Phrases fields:

• A-Z
• a-z
• Space (” “)
• ‘ (apostrophe)

You can also use the following Unicode characters in these fields:

Hebrew character set (he-IL)

For Hebrew dictionary terms, you can use the following characters in the Term value, Sounds Like and Example Phrases fields:

• A-Z
• a-z
• Space (” “)
• ‘ (apostrophe)

Korean character set (ko-KR)

For Korean dictionary terms, the supported characters for the Term value, Sounds Like, and Example Phrases fields are listed in the following PDF file: 

Japanese character set (ja-JP)

For Japanese dictionary terms, the supported characters for the Term value, Sounds Like, and Example Phrases fields are listed in the following PDF file: