“Gen AI is the most powerful tool for creativity that has ever been created. It has the potential to unleash a new era of human innovation.” - Elon Musk
If you are interested in applying to GGI's Impact Fellowship program, you can access our application link here.
1.Introduction
Generative AI has garnered a lot of public attention ever since Chat GPT was released by OpenAI. It promises us a world that is imaginable only in science fiction movies and books - a world where most tasks can be automated and humanoid bots can do creative tasks like research, write sonnets, and produce music while every human is reliant on a personal virtual assistant helping them be hyper-productive. However, the excitement is also mixed with a large and valid set of concerns - such as massive loss of jobs, biased foundation models leading to social bifurcations as well as rise of realistic deepfakes which can cause large-scale misinformation. Hence, it would not be an understatement to say that people as well as the government should work to improve their knowledge and familiarity with Generative AI, understand its underlying potential to enhance employee productivity, improve customer delivery services, and also introduce necessary guardrails to ensure responsible practices of Generative AI.
This white paper delves into the transformative impact Gen AI can have on the Indian economy across various sectors including Retail, Banking and Financial Services, IT service companies, and government delivery services. We explore the existing risks and challenges hindering enterprises and government entities from fully leveraging this potential, while also presenting actionable strategies to overcome these obstacles.
What is Generative AI?
Generative AI is a type of Artificial Intelligence that can produce content such as text, images, sound, video, and even code. Most Generative AI systems are based on Machine Learning algorithms and are trained on a large amount of data scoured from the internet. Some examples of Gen AI systems that have been released in the past couple of years include -
1. Text-to-text :- Chat GPT, Gemini, Anthropic, Mistral, Llama 3.
2. Text to image :- Stable Diffusion, Midjourney, Dall-E, Adobe Firefly
3. Text to audio :- Suno.ai, TTS by OpenAI
4. Text to video :- Sora, RunwayML
5. Text to Code :- Chat GPT, Devin, Devika, Gemini, Github copilot, Claude
Some of the Gen AI systems listed above are also multi-modal since they can interact with the “user” through multiple sensory ways which can be some combination of image, text, audio, video, and code. The ability of Gen AI to understand the input and generate a valuable output in real time for a variety of general use cases is what drives the excitement behind the rise of Gen AI systems. The four major ways in which GenAI can revolutionize business models are by providing personalized customer interactions, accelerating employee productivity, reducing operational costs, and opening new revenue streams. However, there are also challenges that enterprises face to achieve this potential such as skill shortage, lack of available infrastructure for implementation, compute infrastructure, deployment and integration challenges, as well as unclear guidelines on government rules and regulations.
A study by EY shows that India could experience a substantial boost in its GDP over the next seven years (2023-24 to 2029-30) due to Gen AI. The cumulative impact on GDP may range from US$1.2 trillion to US$1.5 trillion, contributing an additional 0.9% to 1.1% in annual CAGR. Feasibility, adoption rate, organized sector’s share and the industry segment’s share in India’s economic activity will determine the level of Gen AI’s impact on specific sectors.
Looking at the massive impact that Gen AI can have on the Indian economy, the EY white paper states that the priorities of the government should be:
Building publicly available and accessible structured and unstructured datasets.
Secure the necessary infrastructure needed for the coßmpute power.
Provide clear guardrails to ensure ethical practices of Generative AI.
Creating Centers of Excellence (COE’s) and providing research funding to institutions as well as industry to reduce the skill gap.
Deploying and testing Gen AI tools for public service delivery systems.
Before moving on to the central role that the government can play in harnessing the economic potential of Gen AI, let’s take a look at the impact Gen AI can have across different sectors of the economy.
2.Impact of Generative AI across sectors
2.1 Banking and Financial Services
The Financial sector serves as the backbone of an economy and Gen AI can result in massive economic growth for this sector. The three primary ways are :
1. Improving customer experience: Virtual agents trained on a bank's central data can be utilized to deliver personalized customer interactions based on the past transaction and query history of that particular individual. This can lead to faster and more effective issue resolution thus improving customer experience.
2. Cost savings: Gen AI can be used in content creation to assist marketing design teams resulting in huge cost savings. Also, virtual agents would eliminate large customer service needs and thus, lead to labor cost savings.
3. Employee productivity: Foundation models can also be used to summarize customer interactions and keep track of sentiment analysis which can help the bank create effective targeted marketing campaigns. Furthermore, LLMs trained on the bank’s central data can also be used to create a copilot to help employees make faster and more effective lending decisions, thus leading to shorter lead times and hence higher throughput and productivity.
In the Indian Banking landscape, banks like HDFC Bank, ICICI Bank, and Axis Bank have been quick to adopt Gen AI. For example, HDFC Bank which is India’s only lender with a market capitalization of over $100 billion, will roll out a private LLM-powered website soon to simplify customer experience whereby, through simple prompts, a customer can quickly access the information they are looking for regarding any product and eventually even their bank statement details. Another example would be Axis Bank which is planning to roll out LLM-based virtual assistants to simplify customer’s banking journey and also launch copilots to empower its employees to improve internal operations and efficiency. It is actively engaging and collaborating with cloud service providers (CSP) and software-as-a-service (SaaS) providers to explore options.
Additionally, training a foundation model for a specific use case enhances its repeatability across different divisions. For example, a customer service LLM that is built and trained centrally on bank data can be reused in the retail, wholesale, and wealth management divisions, thereby multiplying the benefits and effectively managing costs.
An example of how a Gen-AI-powered virtual assistant can improve the customer user journey for query resolution and documentation is shown below:
Despite the promising use cases of Gen AI, companies must also pay close attention to the risks associated with the use and deployment of such systems, such as:
Biases - Training the foundation model without proper guidance can result in biased outputs towards certain demographics, making lending and banking services more difficult and unavailable for certain sections of society. An example where biases can arise due to the use of generative AI is credit decision making because if the training data used to develop the AI model contained historical biases against certain demographic groups, the resulting AI system may make similar unfair and discriminatory lending decisions.
User Privacy - Companies must ensure that customer data is not compromised during output and necessary guardrails are established against data leaks.
Hallucination - Gen AI systems tend to hallucinate many times that is they can create fabricated results and present them with fake authoritative proofs - this can lead to fraud, erroneous financial decisions, and loss of customer trust.
Uncertain AI policy of government - Since Gen AI is still a new and emerging technology, governments around the world haven’t yet established definite regulations. Such uncertainty can cause action paralysis among enterprises and hence companies must adopt responsible AI use cases to be disruption-proof against potential future rules and regulations by the government.
Hence, organizations' functional and technical divisions must work together more closely, guided by the leadership’s vision, to ensure successful Gen AI implementation.
2.2 Retail
Retail is a sector that has embraced technological advancements quite early - from using predictive data analytics to manage supply chain risks and forecasting customer demand to adopting VR/AR for immersive customer experiences, it has usually been an early adopter of new technology. According to consulting reports, current retail investments in AI pegged at US$5 billion, are expected to soar to US$31 billion by 2028. In retail, the areas with the highest impact of Gen AI in the value chain are:
Customer engagement: Gen AI can revolutionize customer engagement with retailers and e-tailers. Gen AI offers dynamic interactions, anticipating customer needs and tailoring responses in real time. It can make the digital customer touchpoints smarter and more intelligent. For example, e-commerce websites can now incorporate vision models for users to trial clothes in real time, visualize furniture within their rooms, redesign their homes, and have personalized query resolutions with chatbots trained on information about their SKUs. As an example, eBay is using pre-trained GPT models as a ShopBot to help enhance the customer shopping experience and create virtual trials.
Unstructured data intelligence: Usually many retailers fail to collect data from customers and hence lose out on a lot of customer insights. However, even the companies that do collect good quality customer data, it is usually unstructured. Gen AI is great at extracting insights such as customer sentiment and even customer preferences from such unstructured data heaps.
Content and creativity: The technology can help predict emerging trends and act as a co-pilot for design teams, ensuring that designs and promotions are always a step ahead. Additionally, it can help create content for targeted marketing campaigns and result in cost savings for the company.
Employee training: Foundation models trained on company data can be used to create a conversation and training chatbot for internal use cases to accelerate the learning process of junior employees and internal issue resolution, leading to faster upskilling of employees thus saving costs and improving productivity.
In the Indian retail landscape, the adoption of GenAI has been swift and effective. Examples include e-commerce websites like Flipkart and Myntra that have adopted vision models to create virtual trials, as well as AR to visualize furniture and 3D objects within our physical space. This results in an immersive shopping experience leading to better customer engagement. Amazon India has deployed Gen AI powered virtual chatbots for query resolution and personalized recommendations. Some retailers like Nykaa and Ajio are planning to implement Gen AI tools as copilots for designers to create effective marketing campaigns as well as for faster product development while others such as Reliance retail are using it for customer sentiment analysis.
According to consulting reports, Gen AI could potentially elevate the retail sector’s profitability by 20% by 2025. The technology does not just reduce overheads but can significantly increase sales through personalized consumer experiences. One of the major challenges for the sector in harnessing this potential is the implementation challenge since a large volume of customers can quickly overwhelm the servers and result in large operational costs. Hence, partnering with external tech providers would be the best way for the industry players to implement the benefits of Gen AI.
3.IT service and Technology companies
India is widely recognized as the IT service provider of the world. However with the advent of Gen AI and code-generating models, one might wonder if the business model of IT service companies might soon be obsolete. However, that is far from the case and according to Bloomberg estimates, the AI focused IT services market opportunity is expected to grow at 30% CAGR over the next decade. There are 3 ways in which Gen AI can deliver impact for IT service companies:
Productivity and internal efficiency: Areas like application development can be significantly improved by reducing lead times by using Gen AI copilot to assist in tasks like creating code documentation, building boilerplate code, and testing. It can also be used on company policies and guidelines to assist and accelerate the learning progression of junior developers. Additionally, it can also be used to translate large codebases from one programming language to another. Thus, Gen AI can be used to automate repetitive and “low-value” tasks thus freeing developers for more complex and creative tasks.
Revenue uplift as a technology partner: Most businesses are unclear of the potential use cases of Gen AI and even if they understand the benefits, the training, deployment, and integration with existing enterprise software is a challenging task. Hence, IT service companies can help other enterprises bridge the skill gap needed to implement and integrate Gen AI tools in their workflow. Gen AI market opportunities for IT services companies include advisory services, cloud and data infrastructure, LLM infrastructure, applications, and custom solutions and services. New revenue opportunities include Gen AI strategy- roadmap, system integration, prompt engineering, and risk management.
Customer experience: Gen AI provides more value to the customers of IT service companies by creating personalized and contextualized virtual agents for fast tech stack query resolution, summarization, and text-to-speech capabilities.
Indian IT service companies have often been quick to adopt changes in the technology world and this has enormously benefited them in the past - from cloud migration to data analytics and now the technological revolution brought by Gen AI. The largest Indian IT service company TCS recently partnered with Google to develop a large portfolio of AI-powered solutions and intellectual assets in the areas of AIOps, Algo Retail, robotics, and smart manufacturing. Additionally, it has also partnered with Amazon AWS to provide end to end service for its customers starting with the creation of tailored AI solutions according to their business needs to training models on company data, and finally, deployment and inference optimization using AWS Bedrock and Sagemaker.
The other Indian IT giant Infosys has adopted Gen AI for a diverse range of uses. Firstly, it is using Gen AI to amplify human potential by building a code assistant that enhances developer productivity in tasks like coding, testing, and documentation. It also created a personalized learning assistant that upskills employees faster, while a sales assistant consolidates collective knowledge for client-facing teams. Secondly, it also uses Gen AI as a knowledge partner trained on years of internal Infosys data to help employees have quick access to valuable information, reducing repetitive work, and enhancing productivity. Additionally, Infosys is creating IP platforms, knowledge assets, and software that it can integrate with client’s existing platforms and databases to provide quick and easy deployment for clients.
Thus, Gen AI can result in massive economic gains for IT service companies since they would play a crucial role in assisting other sectors to train, deploy, and integrate Gen AI services into their business models. However, over-reliance on AI could reduce human oversight and lead to biased or unethical outcomes. Data security and privacy are key concerns along with worries around safeguards in areas like bias, accuracy, and transparency. Gen AI is a major disruptor, and how tech companies navigate its risks and rewards will shape their future relevance in the business world.
4.Government and Public Services
Governments around the world are usually criticized by people for being late adopters of technology and having poor tech services compared to private companies. India has done remarkably well in regards to building a tech stack and some examples include UPI, Digital Aadhar, and ONDC. In regards to Gen AI, the Govt of India has committed to invest in Gen AI to improve citizen engagement, streamline policy drafting, enhance internal efficiency, and establish responsible AI policies.
According to a recent report by BCG, GenAI can unlock US $1.75 trillion in annual productivity gains for governments across the globe by 2033. Additionally, the report stated how citizen surveys have shown that there is a direct relationship between customer experience of digital government services and their trust in the government. Thus, Gen AI offers a great opportunity for governments across the globe to increase trust in their citizens through personalized digital experience and improved efficiency. Some use cases for Gen AI in government services are as follows:
Query resolution - Gen AI models trained on demographic data as well as policies of the government can be employed as virtual agents to help citizens understand criteria for beneficiary schemes, resolve potential queries, provide guidance through processes such as tax filings and immigration applications thus enhancing citizen engagement and personalized service delivery.
Policymaking - Utilizing Gen AI’s ability to synthesize extensive documents, prepare reports, and perform research, policy researchers can better analyze policies and potential loopholes to develop more informed and ethical policies.
Regulatory bodies - Analyzing data to spot trends, irregularities, and patterns can aid in environmental monitoring, financial regulation, and assessing impact, effectiveness and compliance of established regulations.
Internal efficiency and employee productivity - Government officials are usually resource strained and this creates bureaucratic inefficiencies. Gen AI models trained on government policies and citizen data can be used as copilots to assist govt agents in making faster decisions and for outsourcing repetitive tasks such as summarizing documents, translating documents, and even creating effective advertising content for social schemes.
Besides these, there might be additional use cases that emerge as the technology matures and issues like hallucination and biases can be controlled further. However, to integrate Gen AI systems into government delivery services, the government has to start taking appropriate steps now which include identifying appropriate use cases, launching small scale pilot projects, establishing responsible AI guardrails, improving AI literacy among citizens, and fostering innovation by bridging the skill and resource gap.
The Government of India has already started implementing programs and initiatives to develop domestic AI capabilities and increase public awareness. For example, the government has recently launched the Bhashini Incentive under MeitY (the Ministry of Electronics and Information Technology) to incentivize domestic development of Gen AI models and solutions tailored to the multicultural and multilingual environment of India. Additionally, it launched the Mission for Skill development initiative to develop the future workforce needed for training, development and communication of Gen AI services. As a founding member of the Global Member of Artificial Intelligence, the government is looking to establish responsible AI practices and launch programs to help educate the public on the benefits of using Generative AI for both their personal and professional lives. The government has also launched various initiatives such as the National Digital Communications Policy and the Data Center Policy to provide incentives for data centers to develop economical compute infrastructure.
However, the citizens of various countries are divided on whether Gen AI should be used by the government and their top sources of concern are potential job losses and accuracy of results/analysis that might lead to biased and unfair government practices. According to the results of a survey done by BCG, the citizens are most comfortable with the government employing Gen AI services for tasks such as translating documents and websites between languages, providing citizens 24/7 access to information through Gen AI chatbots, implementing copilot for assisting government agents to improve internal efficiency. However, citizens were not comfortable with the use of Gen AI for creating social media content, tracking public sentiments, and assisting in decision making for granting access to public services.
Additionally, as shown in the figure below, the survey done by BCG established how the intricate relationship between Gen AI familiarity, perceived benefits v/s risks, and comfort with government deploying such services creates a flywheel of trust that emphasizes the importance of central initiatives that increase people’s familiarity with Gen AI by building awareness through publicly available courses, building relevant skills through R&D support as well as vocational training workshops, and clear responsible AI guidelines to build citizens trust in government’s use of Gen AI services.
Both the citizens and the governments around the globe have found it difficult to keep track of the rapid pace of progress in Gen AI and that has led to rising uncertainties and threatening concerns. Hence, to build citizen trust in Gen AI use cases and harness the economic potential of Gen AI, the Indian government should:
Establish clear rules and regulations - Many policies and regulations already exist for cybersecurity, data privacy, and intellectual property rights. However, the government must work with relevant stakeholders from industry and academia and update these laws to include the changes brought by Gen AI.
Be Transparent with citizens - The government can achieve transparency by publishing their ML algorithms and if possible, the datasets being used to train their internal models. Also, establishing a centralized Gen AI portal can help citizens keep updated on arenas where the government plans to implement Gen AI tools, their responsible AI practices, and a feedback portal for citizens.
Human-in-the-loop model - The government can implement Gen AI tools by combining humans and AI to adopt the human-in-the-loop model. This ensures fairness, accuracy, and efficiency for complex public services like decision making for welfare schemes. Combining Gen AI with human expertise and morality can greatly enhance trust and effectiveness.
5.Challenges for Implementation of Gen AI
The promise of Gen AI to revolutionize customer interactions, increase employee productivity, decrease cost structure, and create entirely new revenue streams is a boon that can accelerate the economic growth of organizations and countries. However, there are many challenges that organizations face today and hence must be tackled swiftly to harness the untapped potential offered by Gen AI. While many organizations have not yet established clear use cases for Gen AI tools, some top performers have already started scaling their Gen AI practices which can create massive competitive advantages in the foreseeable future. The 5 specific inherent advantages because of which the top performers of Gen AI have scaled effectively are:
Strong link between Gen AI and business value - This can be in the form of financial value (employee productivity, new revenue streams, cost savings) or non-financial value (enhanced customer experience). The top performers of Gen AI have focused on specific targeted applications that can create significant value when scaled rather than following a ‘spray and pray’ approach.
Modern technology infrastructure - The companies that have already adopted predictive AI for data analysis can combine their already existing and modular tech stack with Gen AI for content creation purposes. This can shorten adoption time and lead to synergies between the two types of AI, enhancing value. Top performers like Flipkart and Amazon have already long adopted predictive AI analytics and thus, having an advanced tech stack allows them to implement Gen AI services faster and on a much larger scale than others.
Advanced data strategy - Data is central to the training/fine-tuning of Gen AI models. Hence, companies that have adopted best practices to collect, clean, and store structured and even unstructured datasets would have a significant advantage in building their Gen AI services. However, one must note that bad quality data can also destroy value by creating biased/unsafe content. Example of an Indian company that has significant data advantages is Reliance Jio because of its widely used telecommunication service. The user data Jio has naturally gives it an advantage in the race to build effective LLMs.
4. Leadership vision and transparency- As with any change, the leaders of the companies that are successfully scaling Gen AI for their business play a central role. They have a clear vision to upskill its employees to bridge the skill gap, secure the necessary compute power needed, and most importantly, communicate and ensure full transparency among employees.
5. Adopting Responsible AI practices - Since Gen AI is still a rapidly evolving technology, the rules and regulations around it are still uncertain. The companies that have adopted a hybrid model such as the ‘human-in-the-loop’ model to quality check complex queries, use factual data, and adopt responsible AI practices have been faster and more effective in scaling their Gen AI services.
GenAI is becoming an integral part of business ecosystems today and is already a strong source of value and competitive advantage. However, only 10% of companies have mastered how to scale to create value. Companies that lack the above unfair advantages and resources have encountered various challenges in adopting and implementing Gen AI solutions. Some of these problems and their possible mitigation strategies, especially within the context of the Indian landscape are:
1. Education and awareness - Leaders should familiarize themselves with Gen AI tools and educate themselves on its possible use cases. Leaders can initiate dialogue with industry and academia professionals to better understand how their business models can derive value from implementing Gen AI tools.
2. Prioritize and implement cross-functional teams - Instead of trying to implement Gen AI in all business areas across the organization, leaders should identify the highest impact use case which is also feasible and can be developed as a pilot scale project. For implementation, the organization should create a specialized team that includes data professionals, MLOps Engineers, prompt engineers, and business leaders to optimize speed to implementation.
3. Create partnerships with tech providers - One of the biggest cited challenges by C-suite executives is the skill gap and the challenges associated with identifying the relevant foundation model, fine-tuning, and integration with other existing software such as inventory management tools and CRM platforms. To solve this problem, companies should partner with infrastructure providers for cloud compute services and IT service companies for integration challenges.
4. Governance uncertainty - The lack of rules and regulations can lead to action paralysis for many leaders. Additionally, the implementation of Gen AI tools without proper governance can lead to a lack of customer trust, dissatisfaction among employees, and inefficient operations. Hence, companies must partner with relevant government bodies and academia/industry professionals to understand and educate themselves on responsible and safe AI practices. As an example, Wipro has emerged as a company publicly leading responsible AI practices by emphasizing the importance of data privacy, unbiased decision making, and other ethical considerations of AI implementations.
5. Improve people strategy - One of the top concerns of people in regards to Gen AI is the loss of jobs, and hence companies should focus on increasing familiarity of employees with Gen AI tools, launching programs and workshops for upskilling their employees to use Gen AI services for improved productivity and also to maintain transparency during piloting and implementation of Gen AI tools.
Despite challenges and a lack of familiarity, most organizations are hustling to find ways in which they can integrate Gen AI services into their existing business processes to enhance value - not only directly by creating new revenue streams, saving costs, or enhancing employee productivity but also in indirect ways such as crafting personalized customer interactions and using generative copilot for effective marketing content. The time is ripe for companies to put the above learnings and strategies into practice to reap the future benefits of this transformative new technology.
6.Conclusion
Every revolutionary technology seems magical and sometimes even threatening - the introduction of the alternating current by Tesla was highly criticized by Edison and others as unsafe and life-threatening. As of today, alternating current forms the basis of electrical transmission and is considered a fundamental utility. Another example is the advent of computers, which was initially misconceived as machines that could replace humans. However, in reality, computers have simplified our work, made us more productive, and allowed us to tackle complex and cognitive tasks by taking over repetitive and time-consuming processes.
The recent launch of Chat GPT by OpenAI introduced the world to a new technology that is novel and revolutionary because of its ability to generate creative content and converse in a way that emulates human behavior. As with every revolutionary technology, the rise of Generative AI has raised valid concerns about potential job losses and the authenticity of online content. However, the financial and cultural benefits of the technology can far outweigh the risks provided that the government, industry, and academia professionals work together to establish guardrails and responsible AI practices that ensure the ethical use of this technology.
Meet The Thought Leader
Laboni is a mentor at GGI and is currently working at The Bridgespan Group as a Senior Associate Consultant. She takes interest in socioeconomic development issues, public policy, and equity across different vectors of gender, caste, class, and ability, which in turn fuelled her transition from working at a global bank to the social sector. She is an Urban Fellow from the Indian Institute for Human Settlements, Bangalore and has a bachelor's degree in Economics from St. Stephen's College, University of Delhi.
Meet The Authors (GGI Fellows)
Tushar is a person who is driven by curiosity and fascination to understand more about the world around us. This curiosity is what made him choose physics and currently he is a Physics PhD student at the University of California, Los Angeles. The desire to understand more about the world of business and public policy is what drove him to join the supportive and enriching community of GGI. Interested in creating a tangible impact through Generative AI, he recently started a company that develops innovative SaaS.
If you are interested in applying to GGI's Impact Fellowship program, you can access our application link here.
Commentaires