Rudy Lai

AI is transforming every aspect of investment banking, from front-office activities like trading and advisory services to back-office operations and compliance.

As AI technologies continue to evolve, we can expect to see even more innovative applications in the field of investment banking.

Let's take a closer look.

What is Investment Banking?

Investment banking is a specialized division of banking that primarily deals with the creation of capital for companies, governments, and other entities.

These financial institutions play a crucial role in the global economy by facilitating the flow of capital and helping organizations grow and expand. Investment banks offer a wide range of services, including underwriting, mergers and acquisitions (M&A) advisory, corporate finance advisory, sales and trading, research, and asset management.

The role of investment banks has evolved significantly over the years, adapting to changing market conditions, regulatory environments, and technological advancements. In recent years, the integration of artificial intelligence has been transforming various aspects of investment banking operations, enhancing efficiency, accuracy, and the scope of services offered.

How Do Investment Banks Make Money?

Investment banks generate revenue through various services:

Underwriting new debt and equity securities
Aiding in the sale of securities
Facilitating mergers and acquisitions (M&As)
Providing advisory services on complex financial transactions
Trading securities for their own account or on behalf of clients
Managing assets for institutional clients

The revenue mix can vary significantly among investment banks based on their size, specialization, and market conditions. Large, full-service investment banks typically have more diversified revenue streams, while smaller boutique firms might focus on specific services like M&A advisory.

AI helps investment banks drive business impact in several key ways.

On the revenue side, it enhances deal sourcing by identifying potential M&A targets more efficiently, improves trading strategies through predictive analytics, and enables more personalized client services.
Cost reduction comes from automating routine tasks in both front and back offices, streamlining due diligence processes, and optimizing resource allocation.
For risk management, AI excels at real-time market monitoring, detecting anomalies that might indicate fraud or market manipulation, and running complex scenario analyses for stress testing.

Overall, AI's ability to process vast amounts of data quickly and accurately allows banks to make more informed decisions, operate more efficiently, and stay ahead of market trends.

Who Are the Customers of Investment Banks?

Investment banks serve a broad range of key customers, each with unique financial needs.

Large corporations are major clients, seeking services like capital raising, mergers and acquisitions (M&A) advisory, and strategic financial planning.

Governments and their agencies, including sovereign wealth funds, require assistance with bond issuance and public finance management.

Institutional investors, such as pension funds, mutual funds, and hedge funds, rely on investment banks for executing large trades and accessing new investment opportunities.

High-net-worth individuals (HNWIs) seek specialized wealth management, estate planning, and investment services.

Other important customers include financial institutions like commercial banks, private equity firms and venture capital funds that need deal sourcing and exit strategies, and startups and growth companies seeking IPOs or strategic advice.

Additionally, multinational corporations and real estate developers engage investment banks for cross-border transactions and large-scale property developments.

Top investment banks and how they use AI

How is JPMorgan Chase using AI?

JPMorgan Chase has rolled out the LLM Suite, an AI-powered tool designed to aid in writing, idea generation, and document summarization. Already in use by 50,000 staff, it is now being introduced to the bank’s asset and wealth management division. Mary Erdoes, head of this division, highlighted the importance of AI, stating, “This year, everyone coming in here will have prompt engineering training to get them ready for the AI of the future.”

JPMorgan is aggressively expanding its AI capabilities, employing over 2,000 AI and machine learning experts and data scientists. This push has resulted in more than 400 AI use cases across the bank, including areas such as marketing, fraud detection, and risk management. In his annual letter to shareholders, CEO Jamie Dimon compared AI's potential to transformative inventions like the printing press, emphasizing that AI could be integrated into "every single process" within the bank.

One of the standout AI applications is IndexGPT, a tool that uses OpenAI’s GPT-4 model to create thematic investment baskets. This tool generates a list of keywords related to a theme and uses natural language processing to identify relevant companies, aiming to offer a more refined approach to thematic investing. Rui Fernandes, JPMorgan’s head of markets trading structuring, described IndexGPT as a significant step toward integrating AI across the bank’s index offerings.

Lori Beer, JPMorgan’s global CIO, emphasized the importance of working closely with regulators as the bank pilots its generative AI projects. “It’s about helping regulators understand how we build the generative AI models, how we control them, what are the new vectors of risk,” she said, highlighting the need for early engagement to ensure robust controls and compliance.

How is Goldman Sachs using AI?

Goldman Sachs is rapidly advancing its use of generative AI, particularly in code generation, with a goal to fully deploy its first AI tool to thousands of developers. The firm’s Chief Information Officer, Marco Argenti, has emphasized the company's strategic approach of centralizing all proprietary uses of AI on an internal platform.

This method, while initially slowing down the process, has ultimately accelerated development as the platform matured. This internal platform, known as the GS AI Platform, integrates models like GPT-4 and Google's Gemini and allows for the safe fine-tuning of AI with Goldman’s internal data.

The firm is using this AI platform to enhance efficiency, with its most extensive deployment being a generative AI tool that boosts developer productivity by around 20%. Goldman’s developers are now able to create new applications faster, as the platform allows them to build on existing AI frameworks with built-in safety measures.

This approach ensures that data remains secure and regulatory requirements are met, which is critical in the heavily regulated financial services industry.

Although generative AI remains a small part of Goldman’s technology budget, its potential to reshape operations is significant. The firm is cautious about rushing these technologies into production, prioritizing safety and responsibility.

As the industry navigates the complexities of AI adoption, Goldman Sachs exemplifies a balanced approach, aiming to harness AI's benefits while mitigating its risks.

How is Morgan Stanley using AI?

Morgan Stanley has launched the "AI @ Morgan Stanley Assistant," a generative AI tool developed in collaboration with OpenAI. This tool provides financial advisors and their support staff with quick access to a vast database of around 100,000 research reports and documents. Jeff McMillan, Morgan Stanley's head of analytics, data, and innovation, explained that the AI assistant allows advisors to engage more efficiently with clients by handling routine tasks.

The firm is also piloting another AI tool called "Debrief," which summarizes client meetings, drafts follow-up emails, and integrates notes into Salesforce. Vince Lumia, head of Wealth Management client segments, praised the tool for driving "immense efficiency in an advisor’s day-to-day." McMillan emphasized the AI's capability, stating, "The truth is, this does a better job of taking notes than the average human."

Andy Saperstein, co-president of Morgan Stanley, highlighted the significance of AI, noting that it will "revolutionize client interactions" and "bring new efficiencies to advisor practices." CEO Ted Pick added that AI could save financial advisors "between 10 and 15 hours a week," marking it as "potentially game-changing" for the firm's productivity.

How is Bank of America using AI?

Bank of America has been heavily investing in digital transformation and AI technologies. In 2023, the bank spent $3.8 billion on new technologies and plans to invest a similar amount in 2024, with a specific focus on generative AI projects. CEO Brian Moynihan highlighted that AI is now being used not only for cost-saving but also to enhance customer interactions.

The bank's digital assistant, Erica, launched in 2018, has handled over 2 billion interactions and 800 million inquiries from around 42 million clients. Erica is now integrated more deeply into Bank of America's digital ecosystem, supporting various functions like monitoring spending and managing money transfers. In Q2 2024, Erica served 19.6 million users, facilitating 167 million interactions.

Bank of America’s technology initiatives are supported by a $12 billion annual budget managed by CTIO Aditya Bhasin. The company holds nearly 6,600 patents, with 644 granted in 2023 alone, 28% of which were in information security. Despite these advancements, Moynihan acknowledged that the bank continues to work on "data cleanliness" to fully capitalize on AI capabilities.

How is Citigroup using AI?

By 2024, Citi plans to extend access to AI to the majority of its 40,000 coders, underscoring the banks commitment to integrating AI into its operational fabric.

Stuart Riley, Citigroup’s co-chief information officer, emphasizes that AI is being used to amplify the capabilities of Citigroup’s employees rather than replace them. This sentiment is echoed by Citigroup CEO Jane Fraser, who believes that the risks of not embracing AI far outweigh those associated with its adoption.

Fraser asserts that AI has the potential to revolutionize all bank functions, from coding and customer service to fraud detection and compliance.

The bank's AI strategy, outlined by Shadman Zafar, Citigroup’s co-CIO, revolves around small, incremental changes rather than a single transformational project. Zafar believes that integrating AI into everyday tasks will have the most significant impact on overall efficiency. He also notes that AI’s influence will be far-reaching, potentially changing how employees work for decades to come.

Citigroup employees have pitched over 350 use cases for generative AI, with applications spanning document analysis, compliance, and system modernization. One notable application involved using AI to parse 1,089 pages of new capital rules issued by U.S. federal regulators. The AI technology was able to organize the proposal into manageable sections and generate key takeaways, significantly streamlining the risk and compliance team’s work.

AI is also being used to modernize legacy systems, which are often written in outdated coding languages like COBOL. Zafar highlights that AI can translate these old systems into modern environments, simplifying what were previously massive projects. This modernization is crucial for improving the bank’s ability to roll out new services and make changes quickly.

The bank is also exploring AI’s potential to automate routine tasks in operations, such as clearing trades and validating data for loan processing. These roles often involve repetitive tasks that are prime targets for AI automation. By automating these processes, Citigroup aims to free up staff to focus on more complex, value-added activities.

In customer-facing roles, AI is expected to reduce the mechanical tasks that wealth managers and private bankers currently perform, such as compiling reports for clients. This automation will allow these professionals to spend more time on relationship-building and analyzing data, thereby enhancing the client experience.

As Citigroup continues to implement AI, it is prioritizing risk management and ensuring that appropriate controls are in place. Riley and Fraser both stress the importance of human oversight, even when AI is used to generate code or analyze data. This approach ensures that the bank maintains high standards of accuracy and compliance while leveraging AI’s capabilities.

Citigroup is also using large language models to digest and analyze legislation and regulations in the countries where it operates. This application of AI is particularly valuable in ensuring that the bank remains compliant with various regulatory requirements.

The final phase of Citigroup’s AI strategy will involve integrating AI into the bank’s product design and marketing efforts. Zafar envisions a future where AI helps deliver financial products faster and in more customer-centric ways, closing the gap between customer needs and service delivery.

History of AI in Investment Banking

The integration of Artificial Intelligence (AI) in investment banking has evolved significantly since the 1980s, driven by advancements in computer science and data analytics.

2000s: Algorithmic trading took off, setting the stage for more advanced AI in trading. Data mining became prevalent in market analysis and customer segmentation. ML enhanced risk management and fraud detection. Natural Language Processing (NLP) was used for analyzing financial news. The 2008 financial crisis spurred further AI research for better risk assessment.

2010s: The convergence of big data and AI allowed for more complex and accurate models. Sentiment analysis of social media and news began influencing market predictions. AI became mainstream in investment banking, with major banks establishing dedicated AI research teams. Robo-advisors and AI applications in compliance and regulatory reporting gained traction. Deep learning was applied to complex financial modeling, and AI-driven chatbots emerged for customer service. AI was also used in automated due diligence for M&A processes.

2020s: AI integration deepened, with a focus on explainable and ethical AI. Generative AI models like GPT opened new possibilities in finance, and quantum computing began to be explored for its potential in AI applications.

Key trends include regulatory influence, competitive pressure, technological advancements, and changing client expectations. The future of AI in investment banking will likely see further integration and innovation.

AI Use Cases in Investment Banking

Generative AI is revolutionizing investment banking operations. Here are some key use cases:

Market Analysis & Deal Sourcing: AI can analyze vast amounts of data to identify potential M&A targets and conduct preliminary due diligence. Generative AI can synthesize information from various sources to produce comprehensive market reports and industry analyses.
Risk Management: AI models can predict market trends and assess risk factors more accurately, helping in portfolio management and trading decisions.
Customer Experience: Generative AI can create tailored reports and presentations for clients, enhancing the quality of pitches and client communications.
Compliance: AI can assist in monitoring transactions, detecting anomalies, and ensuring adherence to ever-changing regulatory requirements.
Algorithmic Trading: Advanced AI models can develop and optimize trading strategies, reacting to market changes in real-time.
Document Analysis: AI can quickly review and extract key information from legal documents, term sheets, and contracts.

AI Startups Aimed at Serving Investment Banks

Several AI startups are developing solutions specifically for investment banks:

Alphasense: Provides AI-powered search engines for financial documents
Kensho: Offers AI-driven analytics for investment research
Squirro: Develops AI solutions for relationship intelligence in banking
H2O.ai: Provides an AI platform for various banking applications
Symphony: Offers a secure messaging and collaboration platform with AI capabilities

How is generative AI different from traditional AI in investment banking?

Generative AI can create new content, models, and scenarios, whereas traditional AI primarily focuses on analysis and prediction based on existing data.

What are the potential risks of using AI in investment banking?

Risks include algorithmic biases, over-reliance on AI-generated insights, and potential job displacement. Proper governance and human oversight are crucial.

How can investment banks ensure the ethical use of AI?

Banks should implement robust AI governance frameworks, ensure transparency in AI decision-making processes, and maintain human oversight on critical decisions.

What skills should investment banking professionals develop in the AI era?

Professionals should focus on developing skills in data analysis, AI interpretation, and the ability to combine AI insights with human expertise and judgment.

How can smaller investment banks compete with larger institutions in AI adoption?

Smaller banks can leverage AI-as-a-service solutions, partner with fintech startups, and focus on niche applications where they can develop a competitive advantage.

Key takeaways

AI is transforming investment banking across all areas, from front-office activities like trading and advisory services to back-office operations and compliance. Major banks like JPMorgan Chase, Goldman Sachs, and Morgan Stanley are heavily investing in AI technologies to improve efficiency and offer new services.
Generative AI is becoming increasingly important in investment banking. For example, JPMorgan's IndexGPT uses GPT-4 to create thematic investment baskets, while Morgan Stanley's "AI @ Morgan Stanley Assistant" helps financial advisors quickly access vast amounts of research data.
AI is being used to enhance revenue generation (e.g., improved deal sourcing and trading strategies), reduce costs (e.g., task automation), and manage risks (e.g., real-time market monitoring and fraud detection). It's estimated that AI could save financial advisors 10-15 hours per week.
The integration of AI in investment banking has evolved significantly since the 1980s, with major advancements in areas like algorithmic trading, data mining, natural language processing, and deep learning. The 2020s are seeing a focus on explainable and ethical AI, as well as the exploration of quantum computing for AI applications.
While AI offers numerous benefits, there are also potential risks such as algorithmic biases and over-reliance on AI-generated insights. Investment banks are focusing on implementing robust AI governance frameworks, ensuring transparency in AI decision-making processes, and maintaining human oversight on critical decisions.

AI use cases in investment banking