Zoho unveils proprietary LLM, to be integrated across its product suite

Saas firm Zoho has announced the launch of its proprietary large language model (LLM) Zia LLM. The company said the LLM is trained specifically for business use cases, with a focus on privacy and governance. Embedded across Zoho’s offerings, the LLM helps with a range of tasks such as structured data extraction, summarisation, RAG, and code generation.

Today, Zoho launched three models with 1.3 billion, 2.6 billion, and 7 billion parameters, each separately trained and optimised for contextual applicability. The company said that Zoho will scale the model sizes with the first set of parameter increases by the end of 2025.

Zia LLM has been deployed across Zoho's 16 data centres in the US, India, and Europe. The model is currently testing for internal use cases across Zoho's broad app portfolio and will be available for customer use in the coming months.

According to Zoho, the model has been built entirely in-house, leveraging NVIDIA’s AI-accelerated computing platform under a strategic partnership established last year. The model was trained on 256 NVIDIA GPUs. Zoho has already invested over $100 million in NVIDIA GPUs and AI infrastructure, with an additional $100 million committed. Between June 2024 and June 2025 alone, the company deployed GPUs, including the H100, H200, and L40S, valued at $20 million.

“Our AI investments scale in line with customer adoption, a pattern consistent across the SaaS industry. To avoid overinvestment, we provision resources for a subset of users, then optimise usage to serve a broader customer base. This ensures efficiency and cost-effectiveness. It’s also why we released our 1.2–1.3 billion parameter models first, which are smaller and more efficient, despite the additional development effort they required,” Mani Vembu, CEO of Zoho told TechCircle.

Apart from Zia LLM, Zoho also launched Automatic Speech Recognition models in English and Hindi, a no-code agent builder, Zia Agent Studio, as well as a model context protocol (MCP) server. Zoho also launched over 25 ready-to-deploy Zia Agents, including a few specifically for Indian customers.

Zoho has embedded a suite of AI agents directly into its products, enabling contextual automation across various business functions based on user roles. These include the Customer Service Agent in Zoho Desk, Candidate Screener, which ranks job applicants based on specific criteria; the Deal Analyser, which provides insights on win probability and next steps; and the Revenue Growth Specialist, which identifies upsell and cross-sell opportunities.

For Indian businesses, Zoho has also introduced verification agents for PAN, Voter ID, GSTIN, Udyog Aadhar, and more, supporting use cases such as HR background checks and document verification in financial services. These agents are available through the Zia Agent Marketplace, which will soon allow partners and developers to build and publish their own agents.

“Our key differentiator lies in our native low-code tools, which enable faster time to value, the speed and flexibility of development. By combining low-code with AI, we are also enabling true co-creation, where users can build and refine solutions with AI assistance. Importantly, we keep humans in the loop to verify and guide AI outputs, ensuring accuracy and control. As AI evolves, it will further enhance and strengthen our low-code capabilities,” said Vembu.

With the AI features, subscription prices will remain unchanged, with the additional costs being absorbed by the company. Following a model similar to APIs, users will receive a set amount of free credits, with charges applied only for usage beyond that limit. This strategy will help maintain market competitiveness and broaden customer reach, despite the higher infrastructure investment required for AI, added Vembu.

Notably, Zoho is doubling down on its deep tech strategy. This month, the company acquired Kerala‑based deep-tech startup Asimov Robotics to accelerate innovation in robotics and artificial intelligence. Along with the acquisition, Zoho also inaugurated a new campus in Kottarakkara to expand its R&D footprint.