Bengaluru-based startup Gnani.ai is moving beyond voice AI into enterprise workflow automation and domain-specific models. CEO Ganesh Gopalan said, "It is not enough to just have voice AI offerings; we needed to solve problems." The company added 85 clients recently and will train 1,000 students under a new government initiative.
Gnani.ai, a Bengaluru-based AI startup, is expanding beyond voice AI into enterprise workflow automation and domain-specific models, helping businesses reduce reliance on costly global AI platforms.
The company, which initially focused on voice-first agentic AI, is now helping enterprises build proprietary AI models tailored to their business requirements. Its applications include underwriting and payment reconciliation for banking, financial services and insurance (BFSI) clients, a significant part of its customer base. The shift reflects the growing demand for AI solutions that address specific business problems rather than generic applications.
"It is not enough to just have voice AI offerings; we needed to solve problems," Ganesh Gopalan, co-founder and CEO of Gnani.ai, said. The startup is also developing larger versions of its flagship open-weight reasoning model, Evon v3.3, which was first announced at the India AI Impact Summit earlier this year.
The expansion is beginning to reflect in its customer additions. Gnani.ai has added 85 clients in the past six to eight months, according to Gopalan, with adoption spanning industries. It is also building an overseas presence, with clients in Japan, the Philippines, the UAE and the US.
Gopalan said companies increasingly need to develop their own intellectual property rather than rely entirely on models offered by frontier AI firms, whose costs can make sustained deployment expensive. "My suggestion to companies is to build their own IP. Anyone else relying on Anthropic's AI models consistently will struggle to survive," he said.
The company is also looking to strengthen its talent pipeline. As part of a Karnataka government initiative announced at the Bangalore Tech Summit on Tuesday, Gnani.ai will train 1,000 students initially in areas such as multilingual speech recognition, text-to-speech, language models, prompt engineering, retrieval-augmented generation and multi-agent systems. The broader state initiative, involving Google, Salesforce, Pega Systems and Sarvam, among others, aims to upskill around 64,000 students and professionals.
Gnani.ai plans to offer its voice AI and agentic AI tools to students at subsidised rates, allowing them to work on projects, with the possibility of recruiting some of them.
"Earlier, companies used to prefer hiring fresh graduates, but that world is no more. AI upskilling is very important because engineering students need to be ready to join the workforce directly," Gopalan said.
