# Google Launches ATL Saathi AI App for Teachers in India, Expands Cloud and Security > Source: > Published: 2026-07-14 17:32:56+00:00 **July 14, 2026, (Inside AI) —** In a sweeping move to embed AI into India’s education, enterprise, and security fabric, Google announced a suite of tools and partnerships at its I/O Connect India event in Bengaluru. The centerpiece is **ATL Saathi**, a Gemini-powered desktop app for teachers, alongside expanded cloud infrastructure and open security protocols tailored for the Indian market. The announcements target India’s vast developer ecosystem and regulated sectors, with a clear emphasis on data localization and agentic AI safety. Google DeepMind’s India unit developed ATL Saathi in collaboration with **NITI Aayog’s Atal Innovation Mission (AIM)**. The app curates hands-on learning experiments from a curriculum designed by **Atal Tinkering Labs**. This year, **100 schools** will pilot the tool, with ambitions to scale to **10,000 schools**. The move reflects a broader push by tech giants into India’s education sector, where Microsoft and Amazon have also launched skilling initiatives, though none yet integrate a dedicated AI assistant for teachers at this scale. ## Agentic AI and Data Localization Take Center Stage Google is positioning itself for what it calls the “agentic era,” where AI moves from answering queries to executing tasks. **Preeti Lobana**, Country Manager, Google India, framed the strategy: **“As we drive the shift into the agentic era, where AI moves from answering queries to securely executing tasks, our focus is on providing the underlying infrastructure and guardrails the ecosystem needs to scale safely. Today, we are delivering just that across the board - from flexible on-premise cloud environments for highly regulated sectors, to open security protocols for developers, to localized tools for healthcare and classrooms. We want to ensure that the next wave of Indian innovation is secure, trusted, and built on locally relevant foundations.”** To meet India’s data localization rules, Google will let enterprises and the public sector run **Gemini on Google Distributed Cloud** within Indian data centers, fully disconnected from the public internet. This addresses a key demand from regulated industries like banking and government, which have been slow to adopt cloud AI due to data sovereignty concerns. Google also made **Gemini 3.5 Flash** available via its Enterprise Agent Platform, complying with India’s machine learning processing requirements. The model’s smaller footprint could appeal to startups looking to build cost-efficient AI applications. ## Security and Developer Tools Get an Overhaul Cybersecurity gets a boost with **Sec-Gemini v3**, a specialized agent for threat investigation, now available to testers like **Flipkart**. The tool can reason across complex security data at machine speed, a critical need as agentic AI expands the attack surface. Google also open-sourced **CAPSEM**, a secure runtime environment that isolates AI agents, preventing compromised agents from affecting wider systems. This comes as researchers at **IIT Delhi** and **IIT Madras** work with Google on “Guardian Agents” for early threat detection. For developers, Google introduced the **Device Bound Session Credentials (DBSC)** standard, which makes stolen session cookies useless, and an **Agents-to-Payments** protocol for secure, low-value agent-led transactions under **$100**. These open standards could set a precedent for agentic commerce globally. On the language front, **Gemini Live** now supports **25 Indian languages**, including **Sanskrit**, **Bhojpuri**, and **Maithili**. This broadens access for non-English speakers, though real-world performance in low-resource dialects remains to be seen. In healthcare, **AIIMS Delhi** researchers are using Google’s open **MedGemma** models to build India-specific tools for leprosy and sexual health. Google also announced a free **56-hour** program, AI Research Foundations, with **Nasscom** and **IISC Bangalore** to train developers in fine-tuning large language models. The event, attended by over **1,500 developers**, showcased startups like **Adya.AI** and **PolicyBazaar** using Google’s AI. Google also claimed that **91%** of surveyed Play developers using its AI tools saw revenue gains, and **76%** of adults expressed interest in “vibe coding” with AI. These moves underscore Google’s race to lock in India’s developer community and public sector as competition intensifies from OpenAI and Anthropic, both of which have recently established offices in the country. The long-term impact will hinge on how effectively these tools translate into classroom outcomes and enterprise adoption amid evolving regulations.