Is Generative AI Beginning to Deliver on Its Promise in India?

Generative AI (GenAI) has been heralded as a transformative technology with the potential to revolutionize various sectors, from healthcare and finance to retail and customer service. As we navigate through 2024, India stands at a pivotal juncture in harnessing the power of GenAI. This blog delves into the current landscape of GenAI in India, examining the strides made, the challenges faced, and the road ahead for enterprises and startups alike.

Rapid Innovation in Generative AI

The year 2024 has been marked by frenetic innovation in the GenAI space globally, and India is no exception. A multitude of companies have launched proofs of concept (POCs) and are deploying their first GenAI-powered applications into production. Major tech vendors have introduced their models to the Indian market, offering a variety of options tailored for enterprise users. The startup ecosystem in India has been particularly vibrant, rapidly innovating across diverse domains, including language-specific models that cater to the country’s unique linguistic and cultural landscape.

Advancements in Model Capabilities

The capabilities of GenAI models have significantly evolved, extending beyond text to encompass images, videos, and even robotics integration. For instance, OpenAI's Sora model enables the creation of photorealistic videos from text prompts, opening new avenues in content creation, smart virtual assistants, and virtual reality. Enterprises are also enhancing language models through Retrieval Augmented Generation (RAG), which integrates company-specific data into prompts, thereby improving the relevance and accuracy of AI-generated outputs.

GPU Infrastructure Development

A critical enabler for GenAI’s growth is the availability of Graphics Processing Units (GPUs), which are essential for training and deploying large AI models. Recognizing this, Indian companies like Yotta are developing high-performance computing capabilities within the country. Partnerships with global leaders such as NVIDIA have accelerated the establishment of robust GPU infrastructure, ensuring that enterprises have the necessary resources to scale their GenAI initiatives. Additionally, the Indian government has allocated INR 10,000 crore as part of the IndiaAI Mission to support the development of GPU infrastructure through various models and incentive schemes.

Enterprise Adoption: Progress and Challenges

While innovation is thriving, the translation of GenAI’s potential into tangible enterprise value has been gradual. According to EY's analysis, only about 15% to 20% of POCs by domestic Indian enterprises have transitioned into production, compared to 30% to 40% among Global Capability Centers (GCCs). This discrepancy highlights a cautious approach among traditional Indian enterprises, contrasted by the swifter adoption seen in GCCs.

Key Application Areas

GenAI has found application in several key areas within enterprises:

• Internal Processes: Approximately one-third of use cases focus on utilizing GenAI for executing point tasks with intelligent assistants.
• Marketing Automation: Around 25% of applications are related to automating marketing processes using various GenAI tools.
• Document Intelligence: About 20% of use cases involve document summarization, enterprise knowledge management, and search functionalities.
• Customer-Facing Chatbots: Companies are increasingly deploying GenAI-powered chatbots to enhance customer service and user experience.
• Coding Assistants and Process Automation: GenAI is also being leveraged to build coding assistants and automate internal workflows.

Overcoming Challenges

Despite these promising applications, several challenges impede the widespread adoption of GenAI in Indian enterprises:

• Reliability and Hallucination: GenAI models sometimes generate inaccurate or misleading information, a phenomenon known as hallucination. Techniques like RAG and longer context windows are being employed to mitigate this issue, but it remains a significant concern.
• Data Privacy and Sovereignty: Strict regulations around data privacy and the need to keep data within national borders pose hurdles for enterprises, especially in regulated industries.
• Cost Dynamics: The fluctuating costs associated with GenAI services, particularly large language models (LLMs) like GPT-4, have made investment decisions more complex for Indian enterprises.
• Enterprise Readiness: Leveraging GenAI requires a substantial overhaul of digital and data architectures, which many enterprises are still in the process of undertaking.

Government Initiatives and AI Governance

The Indian government plays a crucial role in fostering an environment conducive to GenAI development while ensuring responsible usage. Through the IndiaAI Mission, the government is investing in foundational aspects of the AI ecosystem, including GPU infrastructure and AI research. Additionally, advisories have been issued to digital platforms to exercise due diligence in handling AI-generated content, emphasizing responsible AI utilization and governance.

Strengthening AI Governance

To enhance the delivery of enterprise value from GenAI, the government and enterprises alike are focusing on:

• AI Governance Frameworks: Establishing clear priorities and policies across various AI use cases.
• Data Security Measures: Enhancing data protection against leaks and ensuring compliance with data sovereignty laws.
• AI Transparency and Reliability: Reassessing approaches to ensure AI systems are transparent, reliable, and free from biases.
• Dedicated AI Leadership: Appointing AI officers to steer AI strategy and governance within organizations.

The Rise of Indic AI

India’s linguistic diversity presents both opportunities and challenges for GenAI. To cater to the vast population engaging with digital interfaces in over 20 local languages, there has been a surge in the development of Indic Large Language Models (LLMs). Initiatives like OpenHathi, Bhashini, and BharatGPT are at the forefront, either fine-tuning existing open-source models or building foundational Indic LLMs from scratch. These efforts are crucial for mass adoption of GenAI, similar to how technologies like UPI and social media have been embraced.

Notable Indic LLM Initiatives

• OpenHathi [Sarvam.ai]
• The Indus Project [Tech Mahindra]
• Ambari Kannada LLM [Cognitive Lab]
• OdiaLlama [OdiaGenAI.org]
• Kannada Llama [Tensoic]
• Bhashini [Government of India]
• Airavata and IndicBART [AI4Bharat]
• Krutrim.ai [Ola]
• BharatGPT [Corover.ai]
• Navarasa [Telugu LLM Labs]
• Pragna [Socket Labs]
• Tamil Llama [Abhishek Balachandran]
• Hanooman [SML – Seetha Mahalakshmi Healthcare]

These models are either wrappers around existing foundational models or are being developed as unique, India-focused solutions. Additionally, global models like Meta’s Llama 3, Google’s Gemini, and OpenAI’s GPT series are increasingly supporting Indian languages, further bridging the gap.

Start-Up Ecosystem: Driving GenAI Forward

The Indian start-up ecosystem has been a significant driver of GenAI adoption, particularly among unicorns. EY India’s analysis reveals that approximately 66% of India’s top 50 unicorns are utilizing AI or GenAI technologies, showcasing a proactive stance compared to traditional players.

Emerging Start-Up Themes

Start-ups in India are focusing on diverse applications of GenAI, categorized into several key themes:

• Marketing and Sales: Companies like VoiceOwl, Pixis AI, and Rephrase AI are revolutionizing marketing automation and sales processes.
• Customer Support: Start-ups such as Haptik , Yellow.ai , and Gupshup up are enhancing customer service through GenAI-powered solutions.
• Media Creation and Content Generation: Beatoven.ai, Visual Dub , and Dubverse are enabling innovative content creation and editing.
• DevOps and Infrastructure (LLMops): Maxim.ai, Portkey, and Truefoundry are optimizing DevOps and infrastructure management with GenAI.
• Sector-Specific Applications: Boltzmann for drug discovery, OnFinance for financial services, and Expertia.ai for HR and recruiting are examples of vertical GenAI applications.

Overcoming Compute Constraints

The availability of compute resources, particularly GPUs, is a critical factor for training and deploying GenAI models. To address this, cloud providers are offering GPUs as part of their infrastructure-as-a-service (IaaS) offerings. Companies like Yotta are developing local GPU clouds, partnering with NVIDIA to enhance their data centers with state-of-the-art AI hardware. Strategic partnerships by Reliance Industries and Tata Group with NVIDIA further bolster India’s GPU infrastructure, ensuring sustained growth in GenAI capabilities.

Notable Enterprise Deployments

Several Indian enterprises have made significant strides in deploying GenAI applications, transitioning from POCs to production environments:

IndiGo (InterGlobe Aviation Ltd) : Enhancing Customer Service

IndiGo has implemented a customized AI chatbot, 6Eskai, leveraging GPT-4 technology. This chatbot manages customer queries, facilitates ticket bookings, and provides personalized recommendations, enhancing the overall customer experience.

Innovaccer : Transforming Healthcare

Innovaccer employs AI to integrate healthcare data from multiple sources, enabling better patient care decisions, improved care coordination, and population health management. Their AI-powered solutions facilitate personalized and data-driven healthcare delivery.

Tata Steel : Producing Green Steel

Tata Steel has partnered with an AI tech platform to utilize GenAI for producing green steel, reducing emissions, and enhancing production quality. AI-powered Digital Twins and advanced analytics optimize operations, offering cloud-based recommendations for sinter and blast furnaces.

Ola : Optimizing Mobility Services

Ola has introduced Krutrim AI, a voice-based AI assistant that enhances rider and driver experiences through natural language processing and machine learning. AI-powered route optimization algorithms improve driver efficiency and provide personalized ride recommendations.

Flipkart : Revolutionizing E-Commerce

Flipkart's knowledge assistant, Flippi, uses GenAI and LLMs to offer personalized product recommendations and enhance shopping efficiency through semantic search and multi-modal search technologies.

HealthifyMe App : Personalized Health Coaching

HealthifyMe has launched Ria 2.0, a multilingual conversational virtual coach that provides personalized health coaching by responding to text and voice commands, tailoring nutrition plans based on user data.

The Future of GenAI in India

As GenAI continues to evolve, India’s ecosystem is poised for significant advancements. The dual-speed approach recommended by experts involves:

1. Rapid Experimentation: Enterprises should continue experimenting with GenAI but in a more structured and guided manner, focusing on POCs that can drive full-scale functional transformation.
2. Building Robust AI Platforms: Simultaneously, there is a need to develop comprehensive AI platforms that integrate GenAI with existing enterprise architectures, automation frameworks, and classical AI methods.

Strategic Recommendations for Enterprises

To maximize the value from GenAI, Indian enterprises should:

• Transition from Ad-Hoc Experiments: Move towards scalable AI programs with a focus on delivering and measuring value from POCs.
• Expand Use Cases: Broaden the scope from specific tasks to comprehensive, agent-based applications that can transform key business functions.
• Leverage Hybrid Models: Utilize both closed-source models like GPT-4 for initial development and open-source options for sustainable growth.
• Enhance Data Security: Strengthen data protection measures to comply with privacy and sovereignty regulations.
• Appoint Dedicated AI Leadership: Establish roles such as AI officers to oversee AI strategy and governance.

Generative AI is gradually transitioning from hype to reality in India, with significant progress in innovation, infrastructure development, and enterprise adoption. While challenges such as data privacy, cost dynamics, and enterprise readiness persist, the collaborative efforts of the government, enterprises, and startups are paving the way for GenAI’s widespread adoption. As India continues to build its AI ecosystem, the promise of GenAI to drive digital transformation and create transformative business value is becoming increasingly tangible.

The journey of GenAI in India is just beginning, and the advancements made in 2024 lay a strong foundation for the future. With sustained innovation, strategic investments, and a focus on responsible AI governance, India is well-positioned to harness the full potential of Generative AI, fostering a new era of digital excellence and economic growth.

#AI #GenerativeAI #IndiaAI #TechInnovation #EnterpriseAI #Startups #IndicAI #DigitalTransformation

LinkedIn News India

LinkedIn News

LinkedIn for Learning

LinkedIn

LinkedIn Talent Solutions