Google’s DeepMind unit in India has introduced a desktop web application that provides teachers with a Gemini-powered assistant trained on curriculum from Atal Tinkering Labs under NITI Aayog’s Atal Innovation Mission. Called ATL Saathi, the tool helps teachers design practical learning experiments for students. It is being introduced in 100 schools nationwide this year, with plans to reach 10,000 schools over time, according to Google. The company also revealed new programs for Indian developers along with AI educational tools, partnerships, and product updates to support local AI firms using Google services while meeting data localization rules. The announcements came at the annual I/O Connect India event in Bengaluru on July 14. Major technology companies see India as a key growth area due to its large internet population and developer community. Google reported paying over Rs 5.3 lakh crore to Indian developers in the Google Play and Android ecosystem in 2024, a 28 percent increase. Surveys showed that over 91 percent of Google Play app developers using Google AI tools saw revenue gains, and 76 percent of adults expressed interest in using such tools for natural language app creation. Preeti Lobana, Country Manager for Google India, noted the focus on infrastructure and safeguards for the agentic era of AI. Dr Manish Gupta of Google DeepMind highlighted India’s adoption of AI across economic sectors. The 2026 India I/O Connect event drew over 1,500 developers, with several startups demonstrating AI solutions built on Google models. Google DeepMind is collaborating with Indian institutions to expand a free 56-hour AI Research Foundations program through partnerships with Nasscom, IISc Bangalore, and AVPN. It is also introducing Sec-Gemini v3 to select government and enterprise testers like Flipkart to address agentic AI security risks. The Gemini 3.5 Flash model is now available to Indian enterprises via the Gemini Enterprise Agent Platform in line with local processing rules. Google will enable Indian organizations to run Gemini on Google Distributed Cloud inside domestic data centers for full data localization. Gemini Live now supports 25 Indian languages and dialects including Sanskrit, Bhojpuri, and Maithili. Researchers at AIIMS Delhi are using MedGemma models for India-specific work on leprosy and reproductive health, and Google is open-sourcing CAPSEM, a secure AI runtime environment, for local developers.
