Zoho bets on Nvidia NeMo to develop its own language models.

Zoho, the renowned email service provider, has revealed its intention to develop customized large language models (LLMs) using Nvidia's AI computing platform. These business-oriented LLMs will benefit from Nvidia's NeMo tools and advanced graphics processing units (GPUs) to enhance the implementation of artificial intelligence in Zoho's services.

The company has made an initial investment of $10 million in artificial intelligence, with plans to double this amount in the next year. This investment, along with collaboration with Nvidia, will enable Zoho to expand and refine its AI capabilities for users of its ManageEngine and Zoho.com platforms, which already serve over 700,000 customers. The development of these LLMs is focused on meeting specific business needs, differentiating them from general-purpose AI tools and facilitating faster and more accurate decision-making for organizations.

Zoho's plans in the field of artificial intelligence are not limited to LLMs; they also include multimodal models that can process visual elements and speech-enabled AI tools tailored for businesses. Currently, the company is evaluating the use of Nvidia TensorRT-LLM to further advance the deployment of its LLMs, achieving a 60% increase in performance and a 35% reduction in latency compared to other open-source frameworks it had tested previously.

The partnership with Nvidia provides Zoho with a more suitable environment to handle a broader range of AI workloads. Nvidia's technology allows for the implementation of features like voice-to-text conversion and the high-performance infrastructure needed. Most importantly, these AI capabilities will be integrated into the complete suite of over 100 tools that Zoho offers, optimizing areas such as sales, marketing, customer support, and accounting.

Unlike larger, general-purpose models, Zoho will focus on developing smaller, specialized language models that specifically address business data needs. This will make Zoho’s AI more accessible to companies regardless of the size of their datasets, enabling a more efficient use of artificial intelligence with specific goals and generating relevant insights for decision-making.

Ramprakash Ramamoorthy, head of AI at Zoho, stated: “Many LLMs in today’s market are designed for consumer use, offering limited value for businesses. At Zoho, our mission is to create LLMs specifically designed for a wide range of business use cases. By owning our own technology stack, covering various business functions, we can integrate the fundamental element that makes AI truly effective: context.”