In a world overflowing with AI hype, CoRover is proving that ‘Made-in-India’ AI is not just solving local challenges but shaping global AI benchmarks. While most AI innovations thrive on high-speed internet and cloud power, CoRover is rewriting the rules by building offline-first, voice-driven AI solutions that reach remote corners of India. From empowering pilgrims at Mahakumbh with AI in low-connectivity zones to creating feature-phone-ready conversational agents through BharatGPT, CoRover is redefining what ‘inclusive AI’ means.
In this exclusive conversation, Ankush Sabharwal, Founder & CEO of CoRover, unpacks how sovereign AI, telephony bots, and conversational commerce are reshaping digital access while setting new benchmarks for India’s AI leadership on the global stage.
CoRover partnered with ONDC to help pilgrims navigate Mahakumbh via AI. What inspired this and how do you measure impact in such a unique, low-connectivity context?
At CoRover, our mission has always been to create Accessible AI for everyone, everywhere, regardless of connectivity, language, or literacy. Our ONDC partnership for Mahakumbh came from the need to provide real-time, context-aware conversational AI support to millions of pilgrims, using BharatGPT.
In low-connectivity situations, success is measured not by vanity metrics, but by actual usage, satisfaction, and impact. This includes the number of pilgrims served, resolution time, reduction in human reliance, multilingual engagement, and feedback from local authorities. The deployment demonstrated how BharatGPT-powered Sovereign AI can promote inclusive digital change on a large scale
With Bharat.GPT's 'Kaisi Jeebh Laplapayi Challenge' campaign, you brought vernacular-first, conversational AI to feature phones with zero data. Is this the tipping point where ‘Make in India’ AI becomes the export model for the world’s next billion users?
Absolutely. The 'Kaisi Jeebh Laplapayi Challenge' was a milestone in Voice-First Conversational GenAI. It showed that India doesn't just consume innovation; we now create AI Assistants that can work on a global scale. We used BharatGPT, sovereign AI, our secure and human-focused GenAI platform, to enable two-way conversations in dialects and regional languages. This worked on feature phones without any data, reaching remote users at a large scale.
This sets the stage for India’s own Make in India AI to become a global standard, especially for the next billion users in emerging markets. They need Conversational AI (Telephony AI) that is low-resource, prioritizes privacy, and is inclusive by design.
From association with IRCTC to UPI, healthcare, insurance and more, CoRover’s AI services span a wide range of sectors. How do you tailor conversational AI for such domain-specific contexts especially when it comes to balancing personalization, security, and scale?
Our Lifecycle-Based Methodology and Composite AI allow us to develop and regularly optimize our Domain-Specific LLMs, which are trained on highly curated, secure, and validated data. Across the board for every industry whether BFSI, travel & tourism, healthcare, defence, banking or insurance-we implement Virtual Assistants (ChatBots, VoiceBots, VideoBots) with industry-specific ontologies, compliance rules, and intent models.
Personalization is made possible through multilingual, multimodal interfaces AI assistants that learn from context without compromising on privacy. We leverage Secure GenAI infrastructure with role-based access, audit trails, and zero data leakage guarantees. Our modular design allows us to maintain speed, scale, and security even in India's diverse ecosystem.
CoRover claims 10x faster deployment and 90%+ accuracy with its proprietary AI. What makes your stack stand out globally and how do you sustain that across formats and languages?
Our proprietary Conversational GenAI architecture built by CoRover is characterized by several novel features that position it at the global level. At its core is BharatGPT, a multilingual, multimodal, and sovereign LLM architecture optimized for real-time AI experiences. This is complemented by end-to-end DataOps and MLOps pipelines to support ingestion, labeling, training, testing, and deployment of models at high speed. We also leverage a hybrid model with both neural and rule-based Natural Language Understanding (NLU) so that new use cases are easily supported without the need to retrain entire models. Any AI platform is as good as relevant data it has, and we have it! Our AI-as-a-Service platform supports more than 100 languages, both Indian and foreign, and supports real-time translation, text-to-speech (TTS), speech-to-text (STT), and voice cloning. Our conversation agents are not just reactive. They are proactive as well. And our Conversational AI agents are not just Question Answer machines. They are of course informational, and advisory and transactional as well, so they help us even do and to end transactions, including payments also in the conversation way. We have deployed Conversational AI Agents for various use cases like pre-sales, sales, and customer support across multiple platforms and channels.
Collectively, these allow us to deploy solutions with AI in hours rather than days/months with an accuracy level of more than 90% in real-world, multilingual, and resource-constrained environments. We have a few critical use cases where we have achieved 100% accuracy also.
How is CoRover helping build the next generation of AI talent especially across India’s regional and underserved communities? Also, what key skills will be critical for India’s workforce to stay relevant and thrive in an AI-first job market?
We actively collaborate with educational institutions, state government departments, and vocational schools to train young people in foundational areas like Conversational AI, Prompt Engineering, AI Lifecycle Management, and Human-AI Interaction Design, mostly in Tier 2 and Tier 3 cities. By using BharatGPT, we empower people to contribute data, develop vernacular AI models, and interact with real-world applications and thus raise AI-literate communities in India. The CoRover’s Campus AI Program (C-CAP) provides end-to-end and inclusive learning in the form of self-paced video modules, interactive sessions for developing AI applications, and hands-on learning with no-code platforms to develop AI agents and assistants. Learners also gain abilities in developing multilingual conversational AI agents, becoming proficient in good prompt engineering, and using leading large language models, including BharatGPT, ChatGPT, Gemini, Llama, and Bhashini. To empower India to play a leadership role in the international ecosystem of AI, it is essential that the workforce gains key capabilities like multimodal AI development (voice, video, and text), domain knowledge matching large language models (LLMs), ethics-based AI practices, and skills in low-code/no-code AI tools. Most importantly, the focus must be on building human-centric AI solutions that not only solve local problems but also meeting global standards.
With video–voice–text bots, AI commerce, and metaverse avatars, what’s your most ambitious vision for CoRover in the next 3–5 years?
Our north star is to serve human kind with technology innovation, so that all can access information and services in their own languages. This is being achieved by providing conversational AI agents to businesses, which enables businesses to automate customer conversations and drive revenue growth, cost savings, and customer satisfaction. With 1.3 billion users and more than 25,000 businesses and developers already using our solutions, we are committed to democratizing technology by making AI assistants as ubiquitous and trustworthy as search engines-only more natural, multimodal, and human-like.
Our vision is to facilitate seamless interactions across different touchpoints by utilizing Telephony AI to facilitate voice communications, Physical AI for smart devices and robots, and enabling all the parts to communicate via conversational interfaces that facilitate effortless human-machine interactions.
In the next 3-5 years, we plan to create metaverse-compatible avatars as multilingual, 3D AI travel guides for sectors such as travel, education, finance, and defence build scalable AI commerce platforms with chat, video, and voice interfaces, and launch offline-first AI assistants driven by Edge AI for underserved areas.
Our technology roadmap encompasses full-stack AI platforms, conversational AI browsers, and operating systems. At the center is BharatGPT, as a Digital Public Infrastructure layer, enabling Ease of Living, Sovereign AI, and equitable digital transformation. We are defining the future of the AI ecosystem, with a focus on accessibility, empowerment, and transformation for everyone.