IBM Cloud Docs
Supported languages

Documentation about IBM Watson® Assistant for IBM Cloud Pak® for Data has moved. For the most up-to-date version, see Supported languages.

Supported languages

Watson Assistant supports individual features to varying degrees per language.

Watson Assistant has classifier models that are designed specifically to support conversational skills in the following languages:

Table 1. Supported languages
Language Language code
Arabic ar
Chinese (Simplified) zh-cn
Chinese (Traditional) zh-tw
Czech cs
Dutch nl
English en-us
French fr
German de
Italian it
Japanese ja
Korean ko
Portuguese (Brazilian) pt-br
Spanish es
Universal* xx

* If you want to support conversations in a language for which Watson Assistant doesn't have a dedicated model, such as Russian, use the Universal language model. For more information, see Adding support for global audiences.

Feature support details

The following tables illustrate the level of language support available for product features.

In the following tables, the level of language and feature support is indicated by these codes:

  • GA: The feature is generally available and supported for this language. Note that features might continue to be updated even after they are generally available.
  • Beta: The feature is supported only as a beta release, and is still undergoing testing before it is made generally available in this language.
  • NA: Indicates that a feature is not available in this language.

Skill support details

Table 2. Skill support details
Language Dialog skill Search skill
English (en) GA GA
Arabic (ar) GA GA
Chinese (Simplified) (zh-cn) GA GA
Chinese (Traditional) (zh-tw) GA GA
Czech (cs) GA GA
Dutch (nl) GA GA
French (fr) GA GA
German (de) GA GA
Italian (it) GA GA
Japanese (ja) GA GA
Korean (ko) GA GA
Portuguese (Brazilian) (pt-br) GA GA
Spanish (es) GA GA
Universal (xx) GA GA

Intent feature support details

Table 3. Intent feature support details
Language Content Catalog Irrelevance detection Enhanced intent detection
English (en) GA GA GA
Arabic (ar) GA (except Covid-19) NA NA
Chinese (Simplified) (zh-cn) NA NA NA
Chinese (Traditional) (zh-tw) NA NA NA
Czech (cs) NA NA NA
Dutch (nl) NA NA NA
French (fr) GA NA NA
German (de) GA (except Covid-19) NA NA
Italian (it) GA (except Covid-19) NA NA
Japanese (ja) GA (except Covid-19) NA NA
Korean (ko) NA NA NA
Portuguese (Brazilian) (pt-br) GA NA NA
Spanish (es) GA NA NA
Universal (xx) NA GA GA

User input processing support details

Table 4. User input processing support details
Language Dictionary-based entity support Fuzzy matching (Misspelling) Fuzzy matching (Stemming, Partial match) Autocorrection
English (en) GA GA GA GA
Arabic (ar) GA GA NA NA
Chinese (Simplified) (zh-cn) GA NA NA NA
Chinese (Traditional) (zh-tw) GA NA NA NA
Czech (cs) GA GA GA NA
Dutch (nl) GA GA NA NA
French (fr) GA GA GA Beta
German (de) GA GA GA NA
Italian (it) GA GA NA NA
Japanese (ja) GA GA NA NA
Korean (ko) GA GA NA NA
Portuguese (Brazilian) (pt-br) GA GA NA NA
Spanish (es) GA GA NA NA
Universal (xx) GA GA NA NA

Entity feature support details

Table 5. Entity feature support details
Language Contextual entities
English (en) GA
Arabic (ar) NA
Chinese (Simplified) (zh-cn) NA
Chinese (Traditional) (zh-tw) NA
Czech (cs) NA
Dutch (nl) NA
French (fr) Beta
German (de) NA
Italian (it) NA
Japanese (ja) NA
Korean (ko) NA
Portuguese (Brazilian) (pt-br) NA
Spanish (es) NA
Universal (xx) NA

System entity feature support details

Table 6. System entity feature support details
Language System entities Legacy system entities
English (en) GA Deprecated
Arabic (ar) GA Deprecated
Chinese (Simplified) (zh-cn) GA Deprecated
Chinese (Traditional) (zh-tw) GA Deprecated
Czech (cs) GA Deprecated
Dutch (nl) GA Deprecated
French (fr) GA Deprecated
German (de) GA Deprecated
Italian (it) GA Deprecated
Japanese (ja) GA Deprecated
Korean (ko) GA Deprecated
Portuguese (Brazilian) (pt-br) GA Deprecated
Spanish (es) GA Deprecated
Universal (xx) GA NA

The Watson Assistant service supports multiple languages as noted, but the tool interface itself (descriptions, labels, etc.) is in English. All supported languages can be input and trained through the English interface.

GB18030 compliance: GB18030 is a Chinese standard that specifies an extended code page for use in the Chinese market. This code page standard is important for the software industry because the China National Information Technology Standardization Technical Committee has mandated that any software application that is released for the Chinese market after September 1, 2001, be enabled for GB18030. The Watson Assistant service supports this encoding, and is certified GB18030-compliant

Changing a skill language

Once a skill has been created, its language cannot be modified. If it is necessary to change the supported language of a skill, you can do so by editing the skill's underlying JSON.

To change the skill language, take the following steps:

  1. Download the skill that you want to edit.

  2. Open the downloaded skill JSON file in a text editor.

  3. Search for the property named language.

    The language property is set to the original language of the skill. For example, the language property is en for an English skill.

  4. Change the value of this property to the language you want to use instead. For example, change it to fr for French or de for German.

  5. Save the changes to the JSON file, and then upload the edited file, overwriting the existing skill.

Configuring bidirectional languages

For bidirectional languages, such as Arabic, you can change your skill preferences.

  1. From your skill tile, click the Actions drop-down menu, and then select Language Preferences.

    This option is only available for skills that are configured to use a bidirectional language.

  2. Select from the following options for your skill:

    • GUI Direction: Specifies the layout direction of elements, such as buttons or menus, in the graphical user interface. Choose LTR (left-to-right) or RTL (right-to-left). If not specified, the tool follows the web browser GUI direction setting.
    • Text Direction: Specifies the direction of typed text. Choose LTR (left-to-right), RTL (right-to-left), or Auto (which automatically chooses the text direction based on your system settings). The None option displays left-to-right text.
    • Numeric Shaping: Specifies which form of numerals to use when presenting regular digits. Choose from Nominal, Arabic-Indic, or Arabic-European. The None option will display Western numerals.
    • Calendar Type: Specifies how you choose filtering dates in the skill UI. Choose Islamic-Civil, Islamic-Tabular, Islamic-Umm al-Qura, or Gregorian.

    This setting is not reflected in the Try it out panel.

    Bidi options

  3. Click the X to close the page. Your changes are saved automatically.

Working with accented characters

In a conversational setting, users might or might not use accents while interacting with the Watson Assistant service. As such, both accented and non-accented versions of words might be treated the same for intent detection and entity recognition.

However for some languages, like Spanish, some accents can alter the meaning of the entity. Thus, for entity detection, although the original entity might implicitly have an accent, your assistant can also match the non-accented version of the same entity, but with a slightly lower confidence score.

For example, for the word "barrió", which has an accent and corresponds to the past tense of the verb "barrer" (to sweep), your assistant can also match the word "barrio" (neighborhood), but with a slightly lower confidence.

The system will provide the highest confidence scores in entities with exact matches. For example, barrio will not be detected if barrió is in the training set; and barrió will not be detected if barrio is in the training set.

You are expected to train the system with the proper characters and accents. For example, if you are expecting barrió as a response, then you should put barrió into the training set.

Although not an accent mark, the same applies to words using, for example, the Spanish letter ñ vs. the letter n, such as "uña" vs. "una". In this case the letter ñ is not simply an n with an accent; it is a unique, Spanish-specific letter.

You can enable fuzzy matching if you think your customers will not use the appropriate accents, or misspell words (including, for example, putting a n instead of a ñ), or you can explicitly include them in the training examples.