Languages Supported for AI+ Capabilities for CFM
Updated
AI+ Builder helps in developing high-quality surveys through generative AI. It facilitates multilingual communication throughout the survey creation process, enabling you to design surveys more effectively with the assistance of AI.
Survey Builder
The languages supported in Survey Builder are:
Spanish
Arabic
Italian
French
German
Portuguese
Romanian
Hindi
Note:
The survey itself is created in the base language, even if the creation experience involves interaction in another language.
Language cannot be changed once the conversation begins.
Translation and interpretation are powered by integrated underlying AI models.
All AI+ Builder functionality relies on LLM language capabilities, and accuracy outside supported languages may vary.
Conversational Surveys
Conversational Surveys use a chat-like interface to deliver surveys. They support the same language capabilities as Survey Builder.
The languages supported in Conversational Survey are:
Spanish
Arabic
Italian
French
German
Portuguese
Hindi
Note:
Survey content is written in English, while the conversation adapts to the respondent's chosen language.
Language switching is not supported mid-session.
Output remains in English for data processing and reporting.
Text Analytics
Sprinklr’s Text Analytics module processes open-ended survey responses to extract meaningful phrases, detect sentiment, and map attributes. It supports multilingual ingestion but varies in output capabilities across components. The languages supported in Text Analytics are:
English
Spanish
Arabic
Italian
French
German
Portuguese
Romanian
Component | Language Handling |
Phrase Extraction | Language-based (multilingual) |
Sentiment | English-only |
Attribute Output | English-only |
Note:
While responses in other languages may be partially processed by the underlying AI (e.g., through OpenAI).
Accuracy, completeness, and consistency may vary for unsupported languages.
Hypothesis Validation
Hypothesis Validation uses LLMs to compare survey findings with external unstructured data (e.g., social, blogs, reviews) to confirm or challenge business hypotheses.
Note:
Survey responses can be submitted in multiple languages, and the model attempts to interpret them using OpenAI’s multilingual capabilities.
Regardless of the input language, the output insight is always generated in English.
English responses yield the most accurate and contextually rich validations.
Survey Insights
The Survey Insights feature generates AI-driven summaries, key drivers, and trend patterns from collected survey responses.
Note:
Multilingual survey responses can be analyzed if supported by the underlying LLM.
The final insights are presented only in English, irrespective of the input language.
For consistent and high-quality insight generation, surveys should be written in English.
Survey Quality
For consistent and high-quality insight generation, surveys should be written in English.
Note:
Surveys written in different languages can be evaluated by leveraging the LLM’s translation and comprehension abilities. However, the output evaluation is always in English.
To ensure detailed feedback and precise quality scoring, it is recommended to use English for survey design.
Response Quality
Sprinklr’s Response Quality evaluation feature is currently supported for the following languages:
English
Spanish
Arabic
Italian
French
German
Portuguese
Romanian
Responses are evaluated based on similar criteria as for response quality, and an appropriate score-based tag (High/Medium/Low/Bot) is assigned to each individual response. For languages outside this list, response quality evaluation will rely on the capabilities of the underlying large language model (LLM). The accuracy and consistency of results in such cases may vary depending on the LLM's proficiency in the respective language.