Zoho Says It Doesn’t Have the Muscle to Compete With ChatGPT

Zoho recently rolled out a range of AI releases, including a proprietary LLM stack and speech recognition models, without ever invoking the usual ChatGPT comparisons. There were no open betas, no claims of general-purpose intelligence, and certainly no talk of disrupting humanity.

Instead, Zoho reiterated what it has quietly believed for years: AI, like software, should serve business needs, not chase internet hype.

While competitors chase benchmark bragging rights and flashy multimodal showcases, Zoho is staying firmly within its lane. 

“We don’t have the muscle to compete with the likes of ChatGPT or Gemini, to be honest,” Ramprakash Ramamoorthy, director of AI research at Zoho, told AIM. “That’s the truth.”

Enterprise Use-Case Only, No Consumer Focus

The company has no plans to open its new Zia LLM models, in 1.3B, 2.6B, and 7B parameter versions, to the public or consumers. These models are not for generic prompting or idle exploration. 

“We want to keep it focused on the enterprise user, and we have a lot of work to be done,” Ramamoorthy said.

Zoho had previously integrated open-source models, such as Llama, Mistral, and DeepSeek, into its stack but found the approach to be insufficient. 

“I think those are very generic; they are built for the consumers, and you are putting that into an enterprise stack,” he added.

The shift to building its own LLMs was as much about long-term control as it was about contextual performance. Ramamoorthy believes LLM providers will eventually evolve into SaaS competitors. Hence, it is essential for Zoho to keep investing in things like this.

AI That Lives Inside the Stack—and Stays There

Zia LLM and Zoho’s new automated speech recognition (ASR) engine are designed to work seamlessly within the company’s own applications—CRM, Desk, Books, Voice, and others. Nothing is routed through third-party APIs. No customer data is being sent to cloud providers for inference.

When asked what differentiates the company from the likes of ElevenLabs, Ramamoorthy said AIM, “This is within the stack, you do not need an extension, you are not sharing your data with another third party. So that is the power of our platform, which makes us a key differentiator, whereas a model service provider might not have the platform; they will have to plug it into something else.”

The LLMs will be bundled into Zoho’s subscription plans, with usage caps by edition, and will not be priced separately. AI agents, including revenue analysers and support bots, will remain free during the adoption phase.

“Agent adoption rates are still poor across the industry, not just in Zoho. It is just a lot of hype around it. So we want customers to see value out of it.”

He continued that the offerings with AI agents and LLMs will be priced only when the customers feel there is value in them.

Ramamoorthy explained that they do not want anyone to be put away from running experiments because of the pricing. “So we are playing a wait-and-watch game here; it is a very conscious decision,” he added.

An Intentional Detour from the AI Arms Race

Zoho’s AI play avoids much of what’s fashionable. The models are relatively small, the training data is a mix of public and proprietary sources, and the focus is purely enterprise. “Nobody is going to plan vacations on a CRM system,” Ramamoorthy said.

Instead, Zoho is building a private backbone for enterprise AI, one where governance, context, and cost control matter more.

Even the speech models, trained in English and Hindi, have only 0.2B parameters, a fraction of the size of OpenAI Whisper V3, yet outperform it in word error rate on standard benchmarks.

Where most players view AI as a product to launch, Zoho sees it as infrastructure to build. And that difference could be potentially gauged with these developments. While others lead with hype, Zoho seems to be leading with integration, embedding AI where it quietly improves outcomes without adding complexity.

It’s not the most headline-grabbing strategy. However, it could be what matters to businesses using platforms like Zoho or those that include it.

In refusing to chase ChatGPT, Zoho might be building something great; only time will tell.

The post Zoho Says It Doesn’t Have the Muscle to Compete With ChatGPT appeared first on Analytics India Magazine.

Generated by RSStT. The copyright belongs to the original author.

Source