Heaptalk, Jakarta — AI-powered customer automation company, Yellow.ai launches its first Large Language Model (LLM) that serves regional languages in Indonesia and offers autonomous customer support in more than 11 regional languages in Indonesia.
With 7 billion parameters, this LLM is trained on an extensive dataset of more than 8.5 billion Indonesian and regional language tokens, including Acehnese, Balinese, Javanese, Minangkabau, and Sundanese. This scheme ensures the relevant outcome contextually, has a high accuracy level, and captures the linguistic nuances of the region, enabling businesses to respond to incoming inquiries up to 90% of the time.
“One effective approach is to ensure that the diversity of Indonesian languages is well represented in the customer service landscape. To achieve this goal, we launched Komodo-7B to support Golden Indonesia 2045,” Rashid Khan, Co-Founder and Chief Product Officer of Yellow.ai, admitted.
The Komodo-7B of Yellow.ai possesses diverse, sophisticated features, including:
- Multi-lingual self-serve: automating customer support with human-like conversations in over 11 regional languages accurately and quickly.
- Advanced Search Capabilities: figuring out search queries entered in regional languages and retrieving relevant information.
- Summarization Tools: summarize long documents, articles, and customer questions to provide faster responses.
- Language-Agnostic Knowledge Base: creating an extensive knowledge base to enable more contextual customer support.
- Colloquial Understanding, responding to customer conversations as in everyday conversations in English, regional languages, and slang.
- Efficient Agent Assist: allowing supporting agents to grant enables support agents to provide solutions from knowledge bases, documentation, and other regional support sources.
- Sentiment Analysis: trained by more than 8.5 million language tokens in Indonesia to acknowledge customers’ tone and sentiment in regional languages and dynamically adjust responses.
LLM Komodo-7B supports understanding the daily conversations of Indonesian natives, a mixture of English and regional dialects. With these advanced features embedded, businesses can provide their responses accurately and more promptly and minimize operational expenses by up to 60%.
“We set up this technology to comprehend customers’ intention, tone, and sentiment in multiple regional languages by yielding response contextually and dynamically, with a customer satisfaction score of 40%,” added Rashid.
Founded in 2016 by Raghu Ravinutala, Rashid Khan, and Jaya Kishore Reddy, Yellow.ai deploys artificial intelligence (AI) to help enterprises reach and engage with customers conversationally and in myriad ways.
In August 2021, this Bengaluru-based company raised US$78.15 million in a Series C funding round, bringing the total investment value up to US$102 million. This AI startup used the funds to fuel international growth across the US, Europe, Latin America, and Asia-Pacific markets. Yellow.ai will also invest further in its product innovation roadmap, deepening multilingual voice bot capabilities, expanding enterprise integrations, and launching a developer marketplace for virtual assistants.
