In an interview with TimesTech, Anurita Das, Founder & CEO of Genovation Technological Solutions, spoke about building Mentis, India’s first cost-effective, privacy-first AI platform. Highlighting its air-gapped, on-premise design, she explained how Mentis ensures data sovereignty, regulatory compliance, and cost efficiency. Anurita also shared insights on agentic AI, enterprise adoption across sectors, and her vision for strengthening India’s AI ecosystem through secure, explainable, and responsible innovation.
Read the full interview here:
TimesTech: Genovation’s flagship platform, Mentis, has been described as a game-changer in the AI space, especially for enterprises concerned about data privacy. What drove you to build a fully air-gapped, on-premise AI system, and how has the market responded so far?
Anurita: As the subcontinent’s first cost-effective, made-in-India AI solution, Mentis is designed to give enterprises complete sovereignty over their data. The idea behind building Mentis as a fully air-gapped, on-premise AI system came from our focus on privacy-first and secure solutions. Unlike cloud-reliant systems, Mentis runs directly on a company’s own servers, giving full control over where it lives and how it operates. In air-gapped mode, the servers are completely isolated from the internet or any external network—like putting data in a vault and cutting all outside connections. This ensures maximum control and significantly reduces risks of data leaks, while also making regulatory compliance easier with frameworks like HIPAA, GDPR, and RBI Guidelines. Mentis goes beyond traditional generative AI by focusing on small, task-specific models (SLMs) that can perform tasks autonomously. By building agentic AI systems with SLMs, it reduces the cost of such iterations by 15x.
As for the market response, we already have a few Fortune 500 companies that are non-Indian entities, and most of our clients today are outside of India or are organizations of foreign origin. We work with clients in various sectors such as manufacturing, retail, healthcare, finance and also with brands in critical sectors including aerospace and defence, namely Airbus Aerospace and DRDO (Defence Research and Development Organisation. We continue to do R&D and have more ideas in the pipeline to expand its capabilities further
TimesTech: In a space dominated by cloud-based giants like OpenAI and Claude, what makes Genovation’s cost-efficient, privacy-first model stand out—and why do you believe this is the future for AI in regulated industries?
Anurita: Genovation takes a different approach by focusing on smaller, task-specific models (SLMs) that can be deployed anywhere, including on-premises and in air-gapped environments. This makes Mentis approximately 15x more cost-effective than commercial AI platforms like OpenAI or Claude, offering one of the first affordable enterprise-grade AI solutions developed in India. Unlike larger models that remain highly compute-intensive and costly to deploy, our system ensures efficiency and adaptability for edge and sensitive applications.
At the same time, Mentis is built with privacy at its core. There are no external API calls, cloud dependencies, or vendor lock-ins—everything runs in a secure, controlled environment. This transparency, combined with encrypted data handling and explainable AI that cites sources and shows raw execution outputs, builds trust. In regulated industries where compliance and data security are critical, this model reduces risks and makes enterprise adoption far more viable.
TimesTech: You’ve spoken about “agentic AI” and its role in transforming how businesses interact with data. How does Mentis use this concept to go beyond traditional generative AI capabilities?
Anurita: Mentis is built as an agentic AI system that goes beyond traditional generative AI by focusing on small, task-specific models (SLMs) that can perform tasks autonomously. Just like humans, when an AI is given tasks, depending on the complexity and steps of failures, the point of failure can increase. Agentic AI can correct itself, but correction means reiterating steps, which traditionally adds up to high costs. By building agentic AI systems with SLMs, Mentis reduces the cost of such iterations by 15x.
At the same time, every output from Mentis is backed by citation, and the AI itself ranks the relevance factors while providing them. This adds an additional layer of judgment, ensuring reliability and transparency.
TimesTech: With over 75% of CISOs reporting rising concerns about AI-related security risks, how does Genovation help enterprises balance innovation with responsibility and compliance?
Anurita: Genovation helps enterprises balance innovation with responsibility and compliance through its privacy-first and explainable AI design. All data is encrypted, particularly at rest, so the system only pulls out the information it requires while thinking, safeguarding sensitive information and maintaining strict privacy standards. By supporting on-premises and air-gapped deployment options, sensitive data never leaves the organization’s controlled environment, reducing risks of unauthorized access.
The explainable AI module ensures that every decision made by the system is transparent—users can see the reasoning, citations, and even raw execution outputs when code is run. This fosters trust and accountability. Additionally, because everything runs without external API calls or cloud dependencies, enterprises avoid vendor lock-ins and meet compliance requirements more easily, including frameworks like HIPAA, GDPR, and RBI Guidelines. This combination of secure deployment, transparency, and encrypted handling ensures innovation without compromising responsibility.
TimesTech: From NASA’s CORE program to founding deep-tech ventures across Asia, your journey is remarkable. How have these experiences shaped your approach to building homegrown, secure AI infrastructure in India?
Anurita: My journey began with early foundations at NASA’s CORE Program, where I focused on AI and machine learning. That experience gave me a strong grounding in research and showed me the possibilities of applying AI to solve complex problems. Later, building and exiting two ventures in Japan and South Korea taught me the importance of hands-on product development and the realities of scaling in different markets.
These experiences directly shaped how I approached Genovation in India. I wanted to create something unique—privacy-first, cost-efficient, and capable of putting India at the front of the agentic AI revolution. By combining rigorous research foundations with practical lessons from building startups abroad, I was determined to design secure, explainable, and enterprise-ready AI infrastructure that could meet both business needs and regulatory requirements here at home.
TimesTech: You’re now part of CII’s ICT subcommittee and actively involved in shaping the policy-tech dialogue. What’s your vision for India’s AI ecosystem over the next 3–5 years, and how do you see Genovation contributing to that future?
Anurita: India’s AI ecosystem is at a pivotal moment. Even though AI seems new, its foundations go back decades, and today we are working through the challenges of privacy, safety, and bias. Over the next 3–5 years, the focus will be on making systems that can protect people’s data while remaining efficient and practical at scale.
At Genovation, we are contributing by building privacy-first, explainable AI systems. Mentis runs in secure, air-gapped environments with no external API calls or cloud dependencies, ensuring data sovereignty. The explainable AI module makes every decision transparent, citing sources and showing raw execution outputs. Alongside this, I actively share insights on practical AI use cases, responsible innovation, sustainability in tech, and aligning emerging technologies with enterprise and policy goals through platforms like CII’s ICT East Subcommittee, an influential advisory body guiding ICT transformation across Eastern India—working closely with government, industry, and academia. I also recently participated in NASSCOM Future Forge 2025, a leading platform for shaping India’s digital future.
