Generative AI – pioneering the next wave in capital markets

  Generative AI has gained prominence in the recent times due to its truly transformative and disruptive potential. The evolution started with rapid advances in machine learning techniques for predictive analytics and insight generation followed by adoption of deep learning models. The models have now evolved into more advanced LLMs (Large language models) which forms the basis for the generative AI models. The LLMs have broken the barriers on language complexity by enabling training on vast amount of data including text, images, and audio for understanding the context, intent etc. across languages, which can result in contextually and semantically correct outputs. Generative AI can now be leveraged across multiple use cases like answer questions based on a knowledge base, summarize topics, write code etc.

The current set of Generative AI applications include ChatGPT, DALL-E, Stable Diffusion, BARD, Midjourney, Deepmind, and others that can process huge organizational data such as text, e-mails, chats, images, video, and audio recordings which can be used to drive business transformations. Some of the benefits include improved customer experience, enhanced productivity, faster product development and reduced costs.

Emerging use cases within capital markets

Major investment and fintech firms have already started experimenting with proof of concepts for various use cases in generative artificial intelligence. Majority of the use cases are focused on improving and transforming customer service, operations, research & insights, and content creation. Generative AI applications provide easy to use APIs for firms to either consume as is or opt to customize the models using proprietary data. These APIs can be seamlessly integrated with the enterprise applications to provide an interconnected platform solution.

Attached picture gives a view on some of the potential use cases for the different lines of business within capital markets based on publicly available information.

  In our view, customer service, content generation and investment research are use cases which majority of the firms are exploring. A brief on the use cases is provided in the subsequent paragraphs.

  Customer service use case includes customer service chatbot that can aid in communication by understanding the intent of the questions, formulate responses and improve the response quality. Data captured from the interactions can also be analyzed for interests and sentiments to pave way for improved customer relationship through hyper personalization. Wealth management firms could leverage the technology to offer personalized investment advice through digital channels, thus enhancing the client experience.

 Relationship managers could also leverage the same for creating personalized marketing campaigns across customer segments, geographies and demographics thus automating the digital sales and marketing. This could potentially increase customer value, conversion and retention over a long period of time. The legal and compliance team could also benefit by generating regulatory and compliance reports thus overcoming the multi format challenges of reporting.

 Generative AI’s extensive data analysis capabilities can be utilized by firms to analyze large volumes of textual analyst reports & recommendations, voice transcripts and data from social media, news, articles etc. to detect patterns, trends, correlations, thus enabling informed investment insights and sound investment decisions.

Current challenges and risks in adopting Generative AI

Although this is a groundbreaking technology, it comes with its own challenges and risks which needs to be effectively managed by the firms for its responsible usage.

Generative AI is at the highest point of the hype cycle. It is important for the firms to explore Generative AI capabilities by identifying a suitable use case which offers business value and helps understand the technology capabilities better. One of the considerations for selecting the use case is data. Since the model outputs are highly data dependent, identifying the right set of data for training, data quality and data security measures needs a closer look.

Challenges remain with leveraging the preexisting models which are already trained on publicly available data sets, as they could potentially contain false and misguided information leading to decision errors.

There are legal and compliance risks pertaining to data privacy and confidentiality, cyber fraud issues and issues relating to explainability of the generated outputs versus human generated ones

How should firms respond to realize the full potential of Generative AI? 

     Generative AI promises to provide significant benefits for the firms. It is important for firms to explore this emerging technology now to gain competitive advantage. Firms need to review their existing innovation portfolio and make generative AI as one of their immediate focus area. Firms need to partner with external providers to bring the best of technology capabilities for improved transformation journey.

The approach is to execute a PoC which would involve identifying business use cases and prioritizing based on validated learning that can be achieved from the use case. One of the approaches could be to explore design thinking and/or lean startup methodologies to achieve maximum benefit. Similar to other AI models, it is important for firms to have a robust AI framework and governance in place with Explainable & trustworthy AI frameworks.

Conclusion 

The global Generative AI market is expected to grow by 34% by 2032 and is expected to increase to USD 165 billion. Firms are increasingly investing in research and development, building POC (proof of concepts), establishing business cases and integration into enterprise platforms. Firms which integrate the capabilities across their front, middle and back-office functions will gain the first mover advantage in the market. As with any emerging technologies, the risks have to be managed with governance and compliance frameworks and ensure careful decisions as it requires significant investments associated with technology infrastructure and workforce.