Zoho Corporation has announced the development of Zia LLM, its own large language model platform aimed at accelerating artificial intelligence integration across its product ecosystem. The company revealed that the model suite consists of three LLMs with configurations of 1.3 billion, 2.6 billion, and 7 billion parameters, each optimized for specific use cases, from structured data extraction to automated code generation. Powered by NVIDIA’s artificial intelligence infrastructure, Zia LLM is currently undergoing internal testing and will be deployed in data centers across Europe, the United States, and India. Public availability will be disclosed in the coming months.
This unveiling is part of Zoho’s broader artificial intelligence strategy, which emphasizes customer data privacy and organizational flexibility by relying exclusively on in-house development. According to Sridhar Iyengar, Zoho’s general manager for Europe, this approach allows the company to democratize access to advanced artificial intelligence technologies while safeguarding sensitive data. The initiative also encompasses proprietary automatic speech recognition models that support English and Hindi, with additional languages planned, as well as the introduction of Zia Agents—pre-built task-specific intelligent agents for domains like sales, customer service, and account management, all hosted via Zoho Marketplace.
Further expanding its artificial intelligence footprint, Zoho has enhanced its code-free Zia Agent Studio, which lets users create smart agents without programming knowledge and provides access to over 700 predefined actions within the Zoho suite. The conversational Ask Zia assistant now supports more sophisticated data analysis, facilitating end-to-end data flow, report generation, and machine learning model creation for business intelligence professionals. Zoho has also introduced a Model Context Protocol (MCP) server, designed to enable seamless integration of its application action library with third-party tools, further underscoring the company’s commitment to platform openness and extensibility within the artificial intelligence landscape.
