Building Disruptive solutions using Generative AI for BFSI industry

15 min readJun 26, 2023

By Dr. Magesh Kasthuri, Swapnil Zarekar, Aishwarya Gupta

1. Introduction

Generative Artificial Intelligence is a class of Machine Learning that has an intelligence to generate new data. Conversational AI, Deep fakes and media/art related generative AI breakthroughs have recently caught people’s attention and imagination with likes of Midjourney, Dall-E and most notably the ChatGPT. Since it’s popularization, the number of product announcements by IT Giants in the field of AI, ML and even Generative AI has skyrocketed, with Google, Microsoft, Meta, GitHub, Palm and Amazon. The key technique behind success of these human like interface into AI is called as Reinforcement Learning with Human Feedback (RLHF) which is an approach that attaches re-evaluated weights to a pretrained model by actual human interference.

It is important to have a controlled approach while exploring potential Generative AI use cases. To identify best suited large language models (LLM) use cases a framework is required. In this document, we have considered use cases for enhancing the customer experience and increasing the productivity in the banking, financial services, and insurance domains. However, Generative AI also finds applications in horizontal functions like product development, product enrichment, operations, sales & marketing, risk & regulations, laws, and human capital.

Figure 1: Generative AI revolutionizing content creation

Generative AI is bringing the paradigm shift in AI. However, below points should be considered while leveraging Generative AI for various business functions

· Cost: Generative AI could be costlier given the core requirement of high computing power

· Data Quality: Generative AI depends on availability of high-quality data to provide accurate responses. This might be a problem for businesses that do not have an easy access to clean data

· Security: The large language models should not consume restricted data and must adhere to the security guidelines for technology and functions

· Explainability & Ethical Concerns: Generative AI might provide biased or unfair responses. It is especially difficult to maintain ethical compliance given the lack of filtering posed by large amounts of data.

Excitement is ahead but a cautious and controlled approach is the key: While cost, data quality, security, and transparency such limitations are not barriers in the journey of exponential growth for Generative AI across industries, they are obstacles and must be worked upon. A controlled approach should be implemented for a scalable Generative AI along with appropriate governance policies. Maintaining data quality, fine-tuning large language models, utilizing alternative green energy sources, and ensuring the data privacy are key leavers for an efficient controlled approach.

It is important to acknowledge that the AI is not inherently unethical or biased, rather the ethical concerns surrounding AI stem from the way it is being trained. Ethical concerns surrounding Generative AI are not unique to it, and every new and emerging technology is subjected to it. Internet, social media, and smartphones were all subjected to a similar treatment during their early days, but were eventually addressed and overpowered through regulations, policies, ethical frameworks, and awareness for the respective technology.

2. Technical Insights to Generative AI

2.1 Outlook into Combinational Technology

While Generative AI offers many benefits, it’s deployment in the industry is of equivalent importance. Leveraging combinations of technologies instead of focusing on a single technology can help in solving critical business problems with scaled implementations. AI can be combined with technologies like Cloud, Bigdata, Internet of Things (IOT) to solve business problems by utilizing values and benefits across the platform.

In the next section of this document, we will be looking at these combinational technologies that go together with Generative AI and can potentially fetch maximum value to businesses. Reduced manual errors and automated workflows are by-products of using combinational technologies. As a testament to this, while AI technologies offer excellent solution in the BFSI domain the industry deployment is rarely standalone. These solutions are often bundled with big data, cloud and/or IOT technologies.

The BFSI sector is constantly seeking new and innovative solutions to enhance customer experiences, optimize operations, and improve decision-making processes. In recent years, the integration of Generative AI with combinational technologies have emerged as a powerful tool to drive digital transformations within BFSI industry. The following section explores the immense impact of combining generative AI with various technologies in delivering advanced solutions.

Internet of Things (IoT) with Generative AI

IoT has revolutionized the BFSI industry by enabling the collection of vast amounts of real-time data from diverse sources. With integration of Generative AI algorithms and IoT devices the banks and insurance companies are gathering valuable insights into their customer behaviour, risk assessment, and fraud detection e.g., smart sensors in insurance policyholder’s house can be utilized to detect potential risks such as fire or theft and generate personalized recommendations or alerts.

Big Data Analytics with Generative AI

The BFSI sector generates enormous volumes of structured and unstructured data, ranging from customer transactions to market trends. Leveraging Generative AI algorithms in combination with big data analytics, financial institutions can derive actionable insights, improve risk assessment models, and detect fraudulent activities more effectively. These technologies enable the identification of patterns and anomalies that may have gone unnoticed using traditional methods.

Blockchain and NFT

Blockchain technology has gained significant attention in the BFSI industry with emergence of Non-Fungible Tokens (NFT), decentralized financial network and Web 3.0 primarily due to Blockchain’s inherent security, transparency, and immutability. Integrating Generative AI with blockchain, financial institutions can enhance security measures, streamline processes, and improve customer trust. For instance, AI-powered smart contracts can be automatically executed based on predefined conditions, reducing the need for intermediaries, and accelerating transaction settlements.

Robotic Process Automation (RPA)

RPA has emerged as a game-changer in automating repetitive tasks and improving operational efficiency in BFSI organizations. Combining Generative AI algorithms with RPA, banks and insurance companies can streamline their manual processes, enhance customer service, and reduce operational costs. AI-powered chatbots and virtual assistants can provide personalized recommendations, handle routine inquiries, and assist in customer onboarding and service processes.

2.2 Generative AI in Cloud platforms

Generative AI based machine learning models deployed over cloud enable faster processing while consuming power more efficiently. This is vital for Generative AI models as they fundamentally require more computational power than traditional processes. Using cloud tech will provide a much clearer ability of control over the cost of any project. Cloud providers including Google, Microsoft, Amazon have already started embedding their Generative AI technology into cloud and offer them as a powerful proposition e.g., Google has come up with embedding generative AI technology in Vertex AI platform which provides integration capabilities with entire ecosystem, Microsoft has launched Azure Open AI (AOAI) which embeds generative AI technology with Azure cloud.

As an example, combining Azure Generative AI with Industry Cloud can offer significant advantages for the BFSI industry such as

Data Integration and Management: Azure Industry Cloud provides industry-specific data models and connectors that can seamlessly integrate with BFSI data sources and systems. By leveraging Azure’s data management capabilities, organizations can aggregate, clean and prepare data for generative AI analysis, ensuring high-quality inputs for AI models.
Generative AI Model Development: Azure provides a range of AI services and tools including Azure Machine Learning used to develop and train Generative AI models. These models can be trained on large datasets to generate meaningful insights and predictions for compliance reporting, risk analysis, fraud detection, and other BFSI use cases.
Scalability and Performance: Azure’s cloud infrastructure offers scalability and high-performance computing capabilities, allowing organizations to process large volumes of data and run computationally intensive generative AI algorithms efficiently. This ensures that BFSI organizations can handle the demands of compliance reporting and analytics effectively.
Security and Compliance: Azure provides robust security and compliance features, including data encryption, access controls, and regulatory certifications. This ensures that sensitive BFSI data used in Generative AI solutions remain protected and compliant with industry regulations, such as GDPR or HIPAA.
Real-time Data Processing: Azure’s stream processing services, such as Azure Stream Analytics, can be combined with generative AI to analyze real-time data streams. This enables organizations to monitor transactions, detect anomalies, and generate compliance reports in near real-time, improving the responsiveness and agility of compliance processes.
Industry-Specific Solutions: Azure Industry Cloud offers pre-built solutions and templates tailored to the BFSI industry. These solutions incorporate Generative AI algorithms to address specific compliance reporting challenges, such as automating regulatory document analysis, anomaly detection, or generating standardized reports.
Collaboration and Knowledge Sharing: Azure’s collaborative features, such as Azure Notebooks and Azure Machine Learning Studio, facilitate teamwork and knowledge sharing among compliance teams, data scientists, and business stakeholders. This collaboration enables organizations to continuously improve the Generative AI models for compliance reporting.

By combining Generative AI with Azure Industry Cloud, BFSI organizations can leverage the power of generative AI to automate compliance reporting, enhance risk management, and improve decision-making while benefiting from industry-specific tools, security measures, and scalability. The combination empowers organizations to harness the potential of AI while aligning with the unique requirements and regulatory landscape of the BFSI industry.

Google offers Generative AI solutions such as DeepMind and TensorFlow, which can be applied to various tasks in the BFSI industry, including compliance reporting. These technologies enable organizations to develop and train generative AI models for data analysis, anomaly detection, and pattern recognition.

Google Cloud Platform provides services such as Google Dataflow and Google Pub/Sub that enable real-time data processing and analytics. BFSI industry can leverage these services in conjunction with Generative AI to monitor transactions, detect anomalies, and generate compliance reports in near real-time. GCP offers tools like AI Platform and Kubeflow Pipelines, which facilitate the development, deployment, and management of machine learning pipelines. These tools can be used to build end-to-end workflows for training and deploying generative AI models for compliance reporting tasks in the BFSI industry.

Following table provides an overview of cloud wise Generative AI features and their applications:

Wipro is one of the preferred partners for all cloud service providers including Microsoft, Google, AWS, and IBM on collaboration with Generative AI technologies to explore and formulate industry specific solution offerings.

2.3 Generative AI in Metaverse solutions

Metaverse provide platform for people to interact with each other and even conduct business in a virtual space. Companies like Meta and Microsoft are investing heavily in the decentralized blockchain technology and implying it in the universe of metaverse, recognizing its potential to be the next big thing in the digital development. Those organizations that effectively use the metaverse will be able to successfully connect, engage with, and incentivize human and machine customers to create new value exchanges, revenue streams, and markets.

It enhances value proposition when combined with Generative AI as the Avatars in Metaverse are then powered by Generative AI capabilities to provide more personalized experience and specific contents that are even better than real conversations. This will boost productivity by enabling dynamic digital interactive Avatars.

2.4 Generative AI in BFSI Industry

According to a report by Research and Markets published in 2020, the global generative AI market size in the BFSI sector was valued at USD 1.18 billion in 2019 and was expected to grow at a compound annual growth rate (CAGR) of 23.2% from 2020 to 2027. BFSI is already in the journey towards digitization by embedding AI in their products and processes. Generative AI would act as an enabler to propel the journey and achieve enhanced customer experience, increased productivity, and bring disruptive innovations.

Wipro has been experimenting with multiple Generative AI use cases across BFSI industry by building rapid prototypes for value assessment and then productionizing some. Wipro sees adoption of Generative AI across Retail/Commercial Banking, Sustainable Finance, Capital Markets/Investment Banking, Wealth/Asset Management, Private Equity, and Insurance sectors. The Wipro Generative AI framework addresses the inherent concerns around Model/Data Bias, Domain & Client Specific Content Moderation. Wipro has embarked the Generative AI journey to provide maximum value to BFSI clients. Below are key potential areas where Generative AI would add value.

Therefore, when it comes to use cases there could be many targeted at increasing productivity/enhancing customer service/brining new business. Below are some of the popular use cases for developing business models using Generative AI solutions in Financial Services industry are:

Fraud Detection: Generative AI can be used to analyze large amounts of data in real-time to identify fraudulent activities. Financial institutions can leverage the technology to create a more secure and reliable system for detecting and preventing fraud.

Credit Scoring: Generative AI can be used to analyze customer data to create a more accurate credit scoring model. This will help financial institutions to make more informed decisions on loan applications and other credit-related activities.

Virtual assistants using Chatbots (e.g., Lending/Credit advisor): Generative AI can be used to develop intelligent chatbots that can provide customers with personalized support and assistance. Chatbots help to reduce the workload of customer service agents and provide faster responses to customer queries.

Risk Assessment: Generative AI can be used to identify potential risks and opportunities in the financial market. This helps financial institution to make more informed investment decisions and reduce their exposure to risk.

Wealth/Portfolio Management: Generative AI can be used to analyze and optimize investment portfolios. This will help financial institutions to make better investment decisions and manage their portfolios more efficiently.

Customer Personalization for core banking customers: Generative AI can be used to analyze customer data and provide personalized recommendations for financial products and services. This will help financial institutions to build stronger customer relationships and increase customer satisfaction.

3. Risk Management in Financial Services

Financial services encounter various risks including credit risk, market risk, liquidity risk, operational risk, and compliance risk. Mitigating or managing these risks is critical for financial services as it may result in heavy financial loss and could have severe consequences. Traditionally financial services industry has been taking various initiatives to help manage the risk including regulations, compliance reports, financial crime proceedings, auditing and they have been using cutting edge technologies to automate the risk management to certain degree.

Compliance reports ensure adherence to multiple regulations of financial services industry. Every financial services firm is a provider of reports whereas multiple regulatory bodies evaluate these reports to identify and curb the potential risks. Over the years there have been advancements in the automation of these reports however, it is still one of the most critical and time-consuming process. Many a times, compliance reports need to be customized per geography, per product, per regulatory changes. e.g., compliance reports for crypto assets are challenging as the industry is still evolving.

Below are some of the key compliance reports in financial services.

4. Risk Management with Generative AI

Generative AI capabilities strengthen the existing risk management measures applied by financial services. Below are some of the key Generative AI initiatives

i) Creation of large volume of synthetic data by using Generative AI to simulate multiple risk scenarios.

ii) Using the synthetic data to retrain the existing machine learning models deployed in risk management functions.

iii) In making risk management models flexible to adapt dynamic financial services trends including frequent regulatory changes, economic conditions.

iv) Enhancing the anomaly detection capabilities of existing models.

Generative AI is a key enabler in financial risk and compliance management function. However, the challenge could be ensuring data security which requires the right Generative AI framework and a controlled implementation.

5. Analyst insights on Risk Management & Compliance reporting

According to Gartner’s Emerging Tech Impact Radar 2023 report, Generative AI solutions comes under productivity revolution theme for which matured time of adoption could be 3–6 years from now.

As per a research report by MarketsandMarkets, the global regulatory compliance market size is projected to reach USD 118.7 billion by 2023, growing at a CAGR of 12.6% from 2018 to 2023. The report highlights that the increasing need for organizations to comply with various regulations and guidelines across different geographies is driving the demand for compliance reporting solutions.

Generative AI solutions are increasingly being used in the BFSI sector for compliance reporting. Additionally, according to a report by Accenture, 76% of surveyed banking executives believe that AI will significantly impact their organizations’ regulatory compliance practices within the next two years.

And a Forrester research suggest that Generative AI solutions can assist in automating compliance reporting processes by analyzing vast amounts of regulatory documents, identifying compliance gaps, and in suggesting necessary actions. The technology can also help in tracking changes in regulations, providing alerts, and generating reports for compliance audits.

Moreover, generative AI solutions can assist in reducing the risk of non-compliance by identifying potential compliance violations and alerting the relevant stakeholders. By leveraging machine learning algorithms, generative AI solutions can detect patterns, anomalies, and deviations in data, enabling organizations to take necessary corrective actions.

Generative AI solutions can also help in enhancing the accuracy and consistency of compliance reporting. The technology can assist in reducing manual errors, ensuring data integrity, and enabling real-time monitoring and reporting.

Generative AI solutions offer significant potential for improving compliance reporting in the BFSI sector. Leveraging these solutions organizations can automate compliance processes, reduce the risk of non-compliance, and improve the accuracy and consistency of compliance reporting. Forrester research expects the AI 2.0 revolution using Generative AI solutions can happen in next 5–10 years.

6. What’s next with Generative AI in BFSI

Generative AI solutions with combinational technologies like Cloud and Metaverse for the BFSI industry will witness significant advancements, enabling more accurate, efficient, and personalized financial services, compliance reporting, risk management, and fraud detection. The integration of emerging technologies and evolving regulatory landscapes will shape the future of generative AI, delivering innovative solutions to meet the industry’s evolving needs.

It’s important to accept the fact that these projections on the future of generative AI may vary, and the actual pace and direction of developments maybe different. The future of generative AI in the BFSI industry will depend on various factors, including technological advancements, regulatory landscape, industry demands, and ethical considerations. Organizations and researchers will continue to drive innovation in generative AI, shaping its future in the BFSI industry. Following table lists future trends and advancements in BFSI industry in next 2–10 years.

7. New age business model with Generative AI solutions

Developing a Generative AI based business model for BFSI industry involves establishing a framework to address business needs (problem statement or use case scenario), data collection, pre-processing data (to prepare data for Generative AI solutions), model training and algorithm development, continuous monitoring and model deployment, business value generation and, finally, return on investment (ROI).

This business model can be applied to various industries, including Banking, Financial Services, Capital Markets, and Insurance to develop customized solutions addressing specific business needs.

New Age Business model services for Generative AI solutions in BFSI

Business model can be driven by the automation capabilities that are offered by Generative AI for any particular use case. Processing, Underwriting & Funding, and Servicing & Disposition are the three broad categories where the use cases can be deployed. Generative AI can touch one or multiple points across the journey.

Above figure depicts the business framework for Generative AI. Once business objectives are defined the corresponding operating model should be designed considering the inherent risk and required compliance measures. It is advised that the framework should be customized for each of the business use case and must be vetted against the security services. While implementing Generative AI the framework, platform, data, application, and governance structure should be evaluated for the final design.

8. Focus area for implementation

Understanding the limitation of current technological landscape and the targeted business objectives is essential. Additionally, focus on the business pain points and how Generative AI enabled solution help in addressing those business pain points is necessary.

It is paramount to proceed with an iteratively growing approach, taking ‘baby steps’, due to the unique and foreign nature of Generative AI. It is advised to begin with 360-degree evaluation of identified use cases, their comparative analysis, their business benefits. Using the business model mapping framework to draft the scope of Generative AI. Data security, compliance, transparency, accuracy, copyrights are critical parameters to be considered and time to market is crucial to be in the race and achieve competitive ROI. Keep in mind that Generative AI might not always be a saleable area but will be a good proposition in boosting innovation capability of an organization and in attracting new clients to new business lines.

Post the advisory workshop and project planning a dedicated focus on implementation is required. It is advised to follow agile implementation approach while keeping an eye on new advancement in Generative AI technology. While simulating business scenarios to validate Generative AI based solution work on the architecture to ensure the Generative AI is customized with required guard rails, embeddings to provide maximum business benefits.

Post implementation, assess the technical and business benefits. Create plan to scale up the implementation and look for other business use cases to embed the Generative AI.