Bhavish Aggarwal’s AI startup, Krutrim, today launched its first major product — an agentic AI assistant named Kruti — positioning it as India’s answer to intelligent, task-driven digital companions. Designed to go well beyond standard chatbot capabilities, Kruti represents a leap toward proactive, multimodal AI tailored for Indian users.
The launch, held on June 12, marked a pivotal moment for India’s growing AI industry. Kruti is being touted as the country’s first fully-fledged agentic AI assistant, built to understand, plan, and execute complex user requests across multiple platforms — including voice, image, text, and mobile interfaces.
A New Breed of AI Assistant
Unlike traditional AI chatbots that wait for commands and offer limited responses, Kruti introduces agentic intelligence, a next-generation AI model that breaks tasks into smaller components, uses specialized internal agents to solve each part, and delivers complete, action-ready solutions. According to Krutrim’s Head of AI, Chandra Khatri, Kruti operates with over 90% real-world task accuracy, offering users a seamless experience that feels closer to a human assistant than a static digital tool.
Built for India: Multilingual, Mobile-First, and Proactive
One of Kruti’s most notable strengths lies in its deep localization. The assistant supports 13 Indian languages, allowing users from diverse regions to interact naturally using both text and voice. Whether it’s Marathi, Tamil, Bengali, or Hindi, Kruti can understand and respond intelligently — a significant advantage in a linguistically diverse nation like India.
Kruti is also designed with a mobile-first approach, making it accessible to the majority of Indian users who primarily use smartphones. It integrates directly with apps like Ola to perform real-world tasks such as booking cabs, ordering food, generating documents, summarizing PDFs, and even creating images — all from a single conversation thread.
How Kruti Works
At its core, Kruti is a multimodal, agentic AI system. This means it can accept various input types, including text, voice, images, and scanned documents. It uses internal “agents” — sub-systems trained for specific functions — to process different elements of a task. For example, if a user sends a bill photo and asks for a summary, one agent extracts the data, another interprets it, and a third formats the response in a human-readable way.
Kruti also includes context memory, enabling it to remember previous interactions and personalize its responses. If you regularly book an Ola cab to work at 9 AM, Kruti will learn your routine and proactively offer to schedule the ride, or suggest your favorite lunch order on weekdays.
Use Case: A Day with Kruti
To understand how Kruti fits into everyday life, imagine a user saying:
“Kruti, schedule my usual Ola cab to the gym, order my regular smoothie, and send the invoice to my health tracking app.”
Kruti could identify the location and time preferences, place a smoothie order from your favorite café, extract the invoice data, and update your fitness records — all without the user opening multiple apps or typing detailed instructions.
This level of autonomy, Krutrim claims, is what distinguishes Kruti from existing AI tools.
Enterprise-Grade Vision and SDK Launch
Beyond consumer use, Krutrim is targeting enterprise adoption with the release of a Kruti SDK (Software Development Kit). This will allow developers and organizations to build domain-specific AI agents using Krutrim’s infrastructure — whether it’s for banking, healthcare, education, or e-commerce.
With the SDK, companies can create tailored assistants capable of handling tasks like automated customer service, intelligent report generation, HR onboarding processes, or even tax filings — all through conversational interfaces.
The enterprise SDK is expected to play a significant role in monetizing the Kruti platform, as Krutrim plans to license its tools to developers and large organizations. This aligns with Bhavish Aggarwal’s broader vision of building an AI ecosystem rooted in Indian needs.
Strategic Investment and AI Independence
Krutrim, already backed by over ₹2,000 crore (~$240 million) in funding, plans to invest an additional ₹10,000 crore into developing homegrown AI technologies over the next year. These include research in speech, translation, computer vision, and advanced LLMs (large language models) tailored to Indian languages and usage.
The company has also indicated plans to host global open-source models like Meta’s Llama 4 on Indian servers at low costs — offering compute access for just ₹7 to ₹17 per million tokens — in a move aimed at democratizing AI development for Indian startups and researchers.
India’s Growing AI Ambitions
With the government actively promoting indigenous AI innovation, Krutrim’s Kruti arrives at a critical time. The country is increasingly focusing on AI sovereignty, data privacy, and economic scalability, and Kruti fits perfectly into that national framework.
Experts from IT firms like Wipro, Infosys, and HCL have already cited 20–35% productivity gains from deploying agentic AI across various sectors. If Kruti’s architecture proves stable and scalable, it could become a core component in India’s digital economy over the next decade.
What’s Next?
Kruti is currently rolling out through integrations with Ola and select developer partners. Public feedback, performance metrics, and enterprise use-cases will determine its future trajectory.
If early adoption succeeds and developers embrace the SDK, Kruti could emerge as a major Indian alternative to global assistants like Google Assistant, Amazon Alexa, or Apple’s Siri — but with a stronger focus on actionable intelligence, local context, and developer customization.
Conclusion
With Kruti, Krutrim isn’t just launching a product — it’s introducing a platform that redefines how Indians can interact with technology. From personalized assistants to enterprise-ready agents, Kruti is designed to do more than talk — it’s built to think, plan, and act.
As Bhavish Aggarwal aptly put it during the launch, “Kruti is the first real step towards the future of AI… built the way Indians live: multilingual, mobile-first, and intuitive.”
Only time will tell how far Kruti will go, but it’s clear that India has just stepped into the agentic AI era — on its own terms.