Here’s What Companies Unveiled at India AI Impact Summit 2026

At the India AI Impact Summit 2026, companies and public institutions unveiled domain-specific AI systems embedded into telecom networks, payment rails, classrooms, healthcare platforms, agricultural supply chains and defence operations. The emphasis shifted away from abstract model benchmarks toward deployment inside high-scale, high-trust environments.

While policy speeches outlined India’s AI ambitions, the product announcements showed where integration is already taking place.

Here’s what our correspondents, Chaitanya and Prabhanu, observed at the summit expo, including the systems demonstrated on the ground and the products launched alongside formal announcements.

I. Foundation Models and Core AI Systems

1. Sarvam AI

Sarvam AI was among the most product-active companies during the summit period, announcing new foundation models while also expanding its deployment ecosystem across speech, vision, agents and infrastructure.

Foundation Models

Sarvam introduced two large language models, including a 105 billion parameter model trained and hosted on India-based data centre infrastructure. A model of this scale places it closer to frontier-sized systems rather than lightweight fine-tuned models.

The company also unveiled Indus, a chat-based application powered by the 105B model. Access to Indus remains limited due to compute capacity constraints.

Speech Systems

Sarvam showcased AI dubbing, speech recognition and text-to-speech systems designed for multilingual Indian language use cases. These tools enable translation and voice generation across languages and are intended for media localisation, transcription and voice-based enterprise or public-sector interfaces.

Vision and Document AI

The company demonstrated document parsing and optical character recognition (OCR) systems capable of extracting text, tables and structured data from Indian language documents. It also presented an AI agent layer that reviews extracted text and corrects errors, allowing documents to be processed more accurately.

Sarvam Kaze

Sarvam unveiled Kaze, AI-powered smartglasses designed to run conversational workflows in a wearable format. The device supports offline functionality, meaning it can operate without continuous internet connectivity.

Agent Infrastructure and Edge Deployment

Sarvam introduced enterprise conversational agent platforms that can connect to backend systems to retrieve information or execute tasks. It also presented multi-agent coordination tools designed to allow several AI agents to work together on structured workflows.

In addition, the company demonstrated on-device AI systems capable of operating in secure or air-gapped environments, meaning systems isolated from the public internet for security reasons.

Compute and State Partnerships

Separately, Sarvam announced partnerships with Odisha and Tamil Nadu to build AI-optimised data centre infrastructure, linking its model development efforts with planned domestic compute expansion.

2. BharatGen – Param2

BharatGen, the government-backed AI initiative, introduced Param2, a multimodal foundation model designed as part of India’s public digital AI infrastructure.

Param2 supports 22 Indian languages and is built to process multiple types of input, including text and images. By positioning it as public infrastructure, the initiative signals an intent to make the model accessible for integration into government services, research institutions and public-sector applications rather than limiting it to proprietary commercial use.

The launch reflects an effort to develop a shared, state-supported AI backbone alongside private-sector model development.

3. Tech Mahindra

Tech Mahindra announced a Hindi-first foundational large language model of approximately 8 billion parameters, tailored primarily for education-focused use cases.

Unlike larger frontier-scale models, this system is positioned as a more targeted deployment model. It is designed to support classroom AI agents, conversational tutoring and structured local-language learning systems.

The focus on Hindi and educational contexts suggests an attempt to build domain-specific AI tools rather than compete directly in general-purpose model development.

4. Gnani.ai

Gnani.ai, an Indian speech technology startup focused on conversational and voice AI solutions, showcased its multilingual voice AI system at the summit. The company’s technology combines speech recognition, synthesis and voice cloning to support a range of use cases across enterprise and public-sector applications.

At the event, Gnani.ai demonstrated capabilities that include text-to-speech synthesis in multiple Indian languages and voice cloning from short audio samples. The system can generate natural-sounding speech and replicate a speaker’s voice with minimal data.

Gnani.ai’s platform targets enterprise workflows such as customer support automation and language-inclusive voice services for government and public interfaces.

It’s worth noting that while several companies announced new models and systems at the summit, OpenAI did not introduce standalone product launches. Instead, it focused on partnerships, including enterprise deployments with Tata Consultancy Services (TCS), infrastructure collaboration through TCS HyperVault, and integrations with platforms such as JioHotstar, MakeMyTrip, IXIGO and Eternal.

OpenAI also expanded its education and enterprise footprint in India, announcing ChatGPT Education licenses and plans to open offices in Mumbai and Bengaluru.

II. Telecom, Infrastructure and Payments AI

5. Jio – AI Ecosystem Expansion

Jio showcased a set of AI applications that it framed as components of a broader “AI stack,” combining sector-focused tools with planned investments in domestic compute and network infrastructure. Rather than announcing a single flagship model, the company presented AI as an embedded layer across multiple verticals.

Jio Arogya AI

Jio Arogya AI is designed as a primary healthcare screening system that can assess symptoms in local languages and generate basic referral recommendations. The tool appears intended for integration into frontline or community health workflows, suggesting use cases in clinics or public health settings rather than as a direct-to-consumer medical application.

Jio AI Homes

Jio AI Homes operates as an AI orchestration layer for connected devices. It is meant to coordinate and automate smart home functions across IoT systems, positioning AI as a control interface within Jio’s broader connectivity ecosystem. The emphasis was on integration with existing digital infrastructure rather than launching a standalone hardware product.

Jio Creator AI 

Jio Creator AI remains under development. The company described it as a tool to generate advertisements and content assets for creators and small businesses. Our reporters at the summit examined the system as part of Jio’s showcase. It was presented as a future-facing capability and has not yet been deployed.

Jio Shiksha Classroom Interface 

The Jio Shiksha Classroom Interface is also in development. It is designed to support structured classroom environments by assisting teachers with lesson delivery and enabling personalised student engagement. Jio positioned it as an institutional support tool for schools rather than a standalone AI tutoring app. It remains an early-stage initiative.

Nation-First AI Stack

Alongside these applications, Jio outlined a “Nation-First AI Stack,” signalling plans to couple sectoral AI systems with domestic cloud, network, and compute investments. The framing suggests an effort to integrate AI services with Jio’s telecom infrastructure rather than operate purely as a model provider.

6. Airtel – AI-Led Digital Infrastructure

Airtel used the summit to showcase how it is embedding AI across its telecom networks, enterprise services and data centre infrastructure, rather than launching a standalone consumer AI product.

AI-Based Fraud Detection Framework

Airtel demonstrated its AI-driven fraud detection system, which operates at the network layer. The system analyses traffic patterns and behavioural signals in real time to identify spam calls, scam attempts and malicious links. By integrating AI directly into telecom infrastructure, the company aims to filter suspicious activity before it reaches end users.

Enterprise Security and Network Intelligence

Beyond consumer spam detection, Airtel highlighted AI-driven enterprise security tools. These include verified business name display features designed to reduce impersonation risks, telecom network APIs intended to support secure digital services and AI-based threat detection aligned with zero-trust security models. The focus was on using AI to strengthen trust within enterprise communication systems.

Cloud and Data Centre Infrastructure

Airtel also outlined investments in AI-ready cloud services and the expansion of its Nxtra data centre network. The company described its infrastructure as optimised to support AI workloads and real-time analytics, while maintaining domestic data residency. It further indicated that AI is being used internally to manage network efficiency and resilience across its distributed data centres.

Overall, Airtel positioned AI as embedded digital infrastructure spanning telecom networks, enterprise security and cloud systems, rather than as a standalone application layer.

7. National Payments Corporation of India – FiMI

The National Payments Corporation of India (NPCI)  announced FiMI, the Finance Model for India.

NPCI described FiMI as a domain-specific language model built for India’s payments ecosystem.

Unlike general-purpose LLMs, NPCI designed FiMI to operate within high-scale, high-trust environments such as Unified Payments Interface and other national payment rails.

NPCI positioned FiMI as optimized for:

• Payments-specific query handling
• Structured financial instruction processing
• High-volume transaction environments
• Compliance-sensitive interactions
• Low-latency responses within payment rails

FiMI reflects vertically integrated AI embedded within the national financial infrastructure rather than layered as an external chatbot interface.

8. PhonePe – Natural Language Search

PhonePe launched an AI-powered natural language search feature built on Microsoft Foundry, Microsoft’s enterprise AI development platform that allows companies to build and deploy generative AI applications using large language models and related tools.

Instead of navigating through menu-based options, users can type or speak instructions such as “Pay Hemanth 20 rupees” or “Recharge FASTag.” The system interprets the intent behind the request and routes the user directly to the relevant transaction or service page.

This replaces traditional menu navigation with intent-based routing, where an AI layer translates conversational commands into specific in-app actions.

By integrating natural language understanding into transaction workflows, PhonePe is shifting from a purely transactional interface toward a conversational layer embedded within its payments infrastructure.

III. Sports & Media

9. Google – AI Cricket Coach

Google demonstrated a Gemini-powered AI cricket coach that uses conversational prompts to guide batting, bowling and fielding techniques. Instead of offering static tutorials, the system allows users to ask questions and receive structured responses in a back-and-forth format, simulating an interactive coaching session.

The demonstration aligns with Google’s broader partnership with the International Cricket Council for the ICC Men’s T20 World Cup 2026. Under that arrangement, Gemini has been designated as the tournament’s Official AI Fan Companion, and Google Pixel as the Official Smartphone.

 VI. Defence and Security AI Systems

10. Indian Army AI Deployments

The Indian Army showcased a set of deployable indigenous AI systems designed for operational readiness, surveillance and decision support in mission environments. Unlike experimental demonstrations, the Army presented these as field-ready tools integrated into defence workflows.

Prakshepan

Prakshepan is described as an AI-based military climatology and disaster prediction system. It analyses environmental and atmospheric data to support operational planning, particularly in regions vulnerable to extreme weather or terrain disruptions.

XFace

XFace functions as a facial recognition platform intended for rapid identity verification in controlled environments. It is positioned for use in security-sensitive deployments where quick authentication is required.

Nabhdrishti

Nabhdrishti is a telemetry and aerial monitoring system that processes real-time data feeds to support surveillance and situational tracking in mission zones.

Driver Fatigue Detection Systems

The Army also demonstrated AI-enabled fatigue monitoring tools designed to assess driver alertness levels, aiming to reduce accidents and operational risks during long deployments.

AI-in-a-Box

AI-in-a-Box is presented as a portable AI toolkit that can be deployed in the field. It allows AI capabilities to operate in constrained or remote environments without reliance on continuous connectivity.

SAM-UN

SAM-UN is a situational awareness module built to integrate multiple data inputs into a unified operational view, assisting commanders in mission planning and execution.

Together, these systems reflect a defence-focused AI stack aimed at embedding machine intelligence directly into field operations rather than treating AI as a standalone experimental layer.

V. Healthcare and Public AI

11. Wadhwani AI – Tuberculosis Detection

Wadhwani AI showcased an early tuberculosis detection system designed to assist frontline health workers in screening for pulmonary TB. The system analyses cough audio patterns along with symptom inputs through a smartphone-based interface to identify individuals who may require further clinical testing.

Rather than functioning as a diagnostic replacement, the tool is positioned as a screening aid intended to support early-stage detection and referral decisions in resource-constrained settings.

12. SATHEE – IIT Kanpur

SATHEE, developed by the Indian Institute of Technology Kanpur, was presented in its upgraded AI-enabled form at the summit. The platform supports competitive exam preparation across national entrance tests such as JEE, NEET, CUET and CLAT.

The system incorporates conversational tutoring, adaptive learning tools and multilingual resources, allowing students to interact with content and resolve doubts through AI-assisted interfaces. Unlike standalone chatbot experiments, SATHEE is positioned as a structured academic support platform integrated into exam preparation workflows.

IV. Agriculture AI

13. Amul – Sarlaben AI Assistant

Amul launched Sarlaben, an artificial intelligence assistant designed for its network of 36 lakh milk producers managing nearly 3 crore cattle. The system is currently being rolled out within Amul’s cooperative network in Gujarat.

Sarlaben is integrated with Amul’s Automatic Milk Collection System and Pashudhan applications, allowing it to draw on operational and livestock-level data across the cooperative database. It provides personalised guidance on cattle health, vaccination schedules, medical treatment, feeding practices and breeding management. The assistant also shares information on government schemes and subsidies available to dairy farmers.

The tool runs on Amul’s existing digital infrastructure, which processes over 200 crore milk procurement transactions annually. Farmers can access it through the Amul Farmer Mobile Application or via phone calls, extending availability beyond smartphone-only users.

What’s Next

The announcements mark only the first stage. The real test will lie in execution, interoperability and long-term integration across sectors.

We will also publish interviews and deeper analyses drawn from our reporting at the summit. Stay tuned for continued coverage and sector-specific deep dives.

Here is a consolidated look at everything we covered during the past week.

Thank you for reading this special newsletter compiling MediaNama’s reporting on the product launches from the India AI Impact Summit 2026.

Also Read:

Support our journalism by subscribing

For You