Issue 21: AI & Gen AI in the Financial Services Industry

Introduction

In February 2024, the approval of the EU AI Act has drawn special attention to the obligations and restrictions of using Artificial Intelligence in the Financial Services Industry. This, coupled with conversations with colleagues in the UK, triggered research in this area and this article. I go into detail below, but let me start with framing the discussing, describing the industry and the solutions.

As I have quoted in other articles, according to a McKinsey report, generative AI could add $2.6 trillion to $4.4 trillion annually in value to the global economy. The banking industry was highlighted as among sectors that could see the biggest impact (as a percentage of their revenues) from generative AI. The technology “could deliver value equal to an additional $200 billion to $340 billion annually if the use cases were fully implemented,” says the report.

NVIDIA's State of AI in Financial Services Report for 2024 reveals that 91% of financial services companies are assessing or already using AI in production and by 2030, AI is expected to boost profits in the banking industry by 20% and could lead to $1 trillion in projected cost savings for the banking and financial sector. According yo the same report, top AI tech in financial services, 26% of the use of AI is in developing/adopting LLMs, 23% in portfolio optimization, 23% in recommender systems and 22% in fraud detection. This is interesting as Fraud Detection was 58% of the use cases for traditional AI.

While all this is great, there is a fine print, that should not be missed. The fine print is that corporate deployment of generative AI in the BFSI industry (Banking, Financial Services and Insurance), is largely nascent. The industry is still primarily using gen AI for cost reduction, rather than revenue growth opportunities. Cost reduction also implies an impact on jobs. According to a recent UBS report, the banking and insurance sectors have some of the greatest proportion of their workforces exposed to potential automation, having a negative impact on employment. The most active use cases revolve around cutting costs by automating low-value, repetitive, time-consuming, tedious jobs, which previously required humans, as the information was largely unstructured and required high engagement with consumers. Now with automation, employees can be freed for more creative work. In fact, in some cases, the tools outperform people.

Background

Let us take a step back to do a primer on the BFSI industry, for those of us that may not be familiar with this space. I do have a fair bit of experience here, having worked on numerous data prep and predictive AI projects for BFSI customers, and leading the release of the Microsoft Cloud for Financial Services. But I am a visual person, so researching some images for this article was awesome for me too.

The financial services industry has the following major segments.

The industry is home to a wide range of services for the management of money, assets, and investments. Here's a brief overview of the main components:

1. Retail Banking: Provides basic banking services to individual consumers, including checking and savings accounts, personal loans, credit cards, and mortgages.
2. Commercial Banking: Offers services to businesses, ranging from small enterprises to large corporations. These services include business loans, treasury and cash management, and sometimes asset management.
3. Investment Banking: Focuses on services for corporations, governments, and other large institutions. These services include underwriting debt and equity offerings, acting as financial advisors, mergers and acquisitions (M&A) advisory services, and facilitating trading.
4. Asset Management: Involves managing investments on behalf of individuals and institutional investors. This includes managing portfolios of stocks, bonds, real estate, and other assets to meet specified investment goals.
5. Wealth Management: Offers more personalized financial and investment advice to high-net-worth individuals, including estate planning, tax planning, and investment management.
6. Insurance Services: Provides risk management through insurance products, including life insurance, property and casualty insurance, health insurance, and liability insurance.
7. Private Equity: Involves investing in private companies (not publicly traded) with the aim of gaining significant influence or control over company operations to increase its value over time.
8. Venture Capital: A subset of private equity focused specifically on investing in startup companies and small businesses with high growth potential.
9. FinTech: Refers to the integration of technology into offerings by financial services companies to improve their use and delivery to consumers. This includes mobile banking, peer-to-peer payment platforms, and robo-advisors.
10. Central Banks: Serve as the national banks that provide financial and banking services for a country's government and commercial banking system, as well as implementing the government's monetary policy and issuing currency.
11. Credit Unions: Member-owned financial cooperatives that provide the same services as retail banks but are typically smaller and focused on serving a community or a group of employees.
12. Payment Processing Companies: Facilitate transactions between two parties, such as merchants and consumers, including credit card processing, payment gateways, and point of sale systems.

The approximate share of the market is below – I did not build a piechart from this because I have found slight discrepancies in these numbers and I did not want a chart to misrepresent. These are approximate numbers, just to provide a framing.

1. Retail Banking: 20%
2. Commercial Banking: 15%
3. Investment Banking: 10%
4. Asset Management: 10%
5. Wealth Management: 5%
6. Insurance Services: 15%
7. Private Equity: 5%
8. Venture Capital: 2%
9. FinTech: 5%
10. Central Banks: N/A (as they do not generate revenue in the traditional sense)
11. Credit Unions: 3%
12. Payment Processing Companies: 10%

?

Within this industry, the use/adoption of AI and gen AI varies by use case of course. This chart demonstrates the use cases well.

Players

As we saw above, BFSI industry has many sub segments and each has its top players.

Here are some top 10 lists as of Jan 31, 2024:

Banks

Rank???? Name??? Market Cap

1??????????? JP Morgan & Chase???????? ?????????????????????????? ????????????$504.07 billion

2??????????? Bank of America???????????? ?????????????????????????? ????????????? $269.15 billion

3??????????? Industrial and Commercial Bank of China????????????? $249.54 billion

4??????????? Wells Fargo??????? ???????????????????????????????????????? ????????????? $182.24 billion

5??????????? China Construction Bank Corp??? ????????????? ???????????? ? $151.76 billion

6??????????? HSBC Holdings?? ???????????????????????????????????????? ???????????? $149.06 billion

7??????????? Royal Bank of Canada???? ?????????????????????????? ????????????? $136.83 billion

8??????????? HDFC Bank Limited???????? ?????????????????????????? ????????????? $131.39 billion

9??????????? Morgan Stanley????????????????????????????????????????? ????????????? $123.05 billion

10???????? Agricultural Bank of China???????????????????????? ??????????? ?? $25.69 billion

?

Insurance companies based:

Rank Name Marketcap

1??????????? United Health Group (UNH) ???????????????????? ????????????? $94.4 billion

2??????????? Berkshire Hathaway Inc. (BRK.B) ???????????? ????????????? $93.2 billion

3??????????? CVS Health Corp Group (CVS)????? ???????????? ????????????? $89.4 billion

4 ?????????? AXA SA (AXAHY) ??????????????????????????????????????? ????????????? $107.6 billion

5??????????? MetLife Inc (MET)????????????????????????????????????? ????????????? $71.6 billion

6??????????? Prudential PLC (PUK)???????????????????????????????? ????????????? $63.5 billion

7??????????? Ping An Insurance China ?????????????????????????? ????????????? $156.2 billion

8??????????? Allianz (ALIZY)???????????????????????????????????????????? ????????????? $26.5 billion

9??????????? Humana Inc. (HUM)?????????????????????????????????? ????????????? $26.4 billion

10???????? Axa Sa (AXAHY)????????????????????????????????????????? ????????????? $22.5 billion

?

Asset Management companies:

Rank???? Firm/company?? ????????????? AUM (Assets under management)

1??????????? BlackRock?????????? ??????????????????????????????????????? ????????????? $9,090 billion

2??????????? Vanguard Group????????????? ?????????????????????????? ????????????? $7,600 billion

3??????????? Fidelity Investments?????? ?????????????????????????? ????????????? $4,240 billion

4??????????? State Street Global Advisors??????? ????????????? ????????????? $3,600 billion

5??????????? Morgan Stanley ???????????????????????????????????????? ????????????? $3,131 billion

6??????????? JPMorgan Chase????????????? ?????????????????????????? ????????????? $3,006 billion

7??????????? Goldman Sachs? ???????????????????????????????????????? ????????????? $2,672 billion

8??????????? Crédit Agricole? ???????????????????????????????????????? ????????????? $2,660 billion

9??????????? Allianz?? ?????????????????????????????????????????????????????? ????????????? $2,364 billion

10???????? United States Capital Group????????????????????? ????????????? $2,300 billion

Market Opportunity

According to Market Research, the opportunity in BFSI is?expected to soar from a value of USD 1,085 million in 2023 to USD 9475 million by 2032, the market for Generative AI in financial services is witnessing an unprecedented compound annual growth rate (CAGR) of?28.1%.

Generative AI's capacity to sift through large data volumes to identify patterns and anomalies empowers financial institutions to make more informed decisions and effectively manage risks. Also, the surge in mobile banking has increased the need for Gen AI-powered chatbots and virtual assistants, offering quick and personalized customer service while ensuring the security of financial information.

Global footprint

The US is leading in the growth in the sector, followed by Europe and Asia Pacific regions - this is largely obvious, due to the availability of data, solutions and talent.

Tech developments

There are many subsegments, which I go into further below, however, at a 30K feet altitude, the leading areas benefiting from gen AI in Financial Services are:

Customer service

• Automated personalized responses
• Enhanced advisory services

Risk management

• Improved fraud detection

Operational efficiency

• Accelerating claims processing
• Streamlining loan approvals
• Automating compliance and underwriting processes

Brief history

Again given the vastness and variety of each segment, it is difficult to draw one picture of how AI impacted the industry. However I thought I’d look for an infographic with familiar events, that most of us readers would be able to relate to. So this image really resonated with me, as it started with the launch of Apple Pay, something I use frequently, if not everyday, and ends with Amazon Go, also something I frequently use.

Solutions

Generative AI in financial services industry is poised to bring transformative solutions across various domains.

Here are some of the top solutions, mapping to the categories above but in some more detail:

In each segment, and in each region, the use of AI vs gen AI and its potential to generate growth value varies. Here is a view of the projected value created by traditional or gen AI.

I found this awesome chart from BCG 波士顿谘询公司 that shows in greater detail where the specific use cases and solutions are. The green colored functions are the three covered up here, and it is obvious that those are the popular areas.

Active implementations

Here are some real use cases being implemented by some leading financial institutions.

• J.P. Morgan Chase- Developed OmniAI platform for AI applications at scale- Utilizes AI for fraud detection, employee and client assistance, and customer response- Introduced AI-powered virtual assistant for corporate clients
• Capital One- Delivers real-time automated financial services using AI and ML
• Royal Bank of Canada- Leveraging pattern-spotting AI software for back-office functions and fraud detection
• Wells Fargo- Proactively using AI technologies like machine learning, deep learning, and chatbots for business processes
• UBS (Switzerland)- Implementing AI solutions for credit decisions and code generation
• Commonwealth Bank- Training AI software for managing back-office functions and client interactions
• Citibank- Harnessing AI and ML technologies for document digitization, customer service chatbots, and fraud detection
• Barclays- Deploying cognitive reasoning, machine learning, and RPA to revolutionize transaction banking

Impact on jobs

Autonomous Research estimates that 1.2M people working in banking and lending will be replaced by AI solutions by 2030, and that made me wonder what type of jobs would be impacted the most. I found this great chart to show the proportion of the impact. It is important to keep in mind that new jobs will be created, therefore there will be reskilling and repurposing of labor.

?

Startups

Artificial intelligence in the banking, financial services and insurance sector has given customers a completely new experience. Customer have demands for a safe, digital platform where they can access, save and invest their money has been met. As the opportunities for AI and gen AI continue to expand, we need new startups and investors alike. The pace of innovation is exciting and the solutions are transformative to the industry. Here is a chart that summarizes the startup activities.

Companies on the market map fall into 9 main categories:

• Credit Scoring / Direct Lending: Companies in this category use AI for robust credit scoring and lending applications. Notable companies include?Affirm? and?ZestFinance .
• Assistants / Personal Finance: Companies in this category rely on an AI chat bot and mobile app assistant applications to monitor personal finances. Notable companies include?Digit? and?Kasisto .
• Quantitative & Asset Management: Companies in this category employ AI algorithmic trading and investment strategies or tools. Notable companies include?Sentient Technologies? and?Numerai .
• Insurance: Companies in this category use AI to quote and insure. Notable companies include?Lemonade? and?Cape Analytics .
• Market Research / Sentiment Analysis: Companies in this category use AI to efficiently research and measure sentiment. Notable companies include?Dataminr ?and?Alphasense .
• Debt Collection: Companies in this category use AI to improve creditor collection of outstanding debt through personalized and automated communication.?TrueAccord ?and?CollectAI ?are included in this list.
• Business Finance & Expense Reporting: Companies in this category use AI to improve basic business accounting, including expense reporting. Notable companies include?AppZen ?and?Zeitgold .
• General Purpose / Predictive Analytics: Companies in this category use AI for general purpose semantic and natural language applications as well as broadly applied predictive analytics.?Notable companies include?Opera Solutions ?and?Kensho Technologies .
• Regulatory, Compliance, & Fraud Detection: Companies in this category use AI to detect?fraudulent and abnormal financial behavior, and/or use AI to improve general regulatory compliance matters and workflows. Notable companies include?Trifacta? and?Digital Reasoning Systems .Below is a table of the companies included in the market map.

The EU AI Act and implications for the BFSI industry

On?9 December 2023, European Parliament negotiators and the Council presidency agreed on the world's first-ever comprehensive legal framework on Artificial Intelligence; the European Union Artificial Intelligence Act (the "EU AI Act"). On 2 February 2024, the EU's Artificial Intelligence Act was unanimously approved by the Council of EU Ministers. This is a major step taken by the EU, as AI Act has previously been plagued by delays. The AI Act focuses on a risk-based approach and aims to foster innovation while ensuring the protection of fundamental rights.

Here's a summary of the key points covered:

Risk-Based Approach

The AI Act classifies AI systems based on their risk level to public health, safety, and fundamental rights, among other factors. Systems are categorized into low-risk and high-risk applications, with the latter subject to more stringent requirements. These include Fundamental Rights Impact Assessments, Conformity Assessments, and mandatory registration in an EU database. The Act aims to ensure that higher-risk AI systems are subject to more rigorous governance to mitigate potential harms.

Prohibited AI Systems

Certain AI applications are banned outright due to their perceived threat to people's fundamental rights. These include biometric categorization systems that identify sensitive characteristics, untargeted scraping for facial recognition databases, emotion recognition in workplaces and schools, social scoring systems, AI that manipulates human behavior, and AI exploiting vulnerabilities of specific groups.

GPAI Systems and Foundation Models

Special rules are set for general-purpose AI systems (GPAIs) to ensure transparency and accountability, particularly for high-impact models that might pose systemic risks. These rules cover technical documentation, compliance with copyright law, and increased transparency about training data.

Promoting Innovation

The Act encourages "regulatory sandboxes" and real-world testing environments facilitated by national authorities, aiming to balance innovation with regulation. This approach allows for the development and refinement of AI technologies in a controlled environment before they are introduced to the market.

AI Regulatory Framework

The establishment of new administrative structures, including an AI Office within the Commission, an independent expert scientific panel, an AI Board comprising EU member states' representatives, and an advisory forum for stakeholders, is a notable aspect of the Act. These bodies are designed to oversee AI development, enforce regulations, and foster standards and best practices across the EU.

Penalties for Non-Compliance

The AI Act stipulates substantial fines for violations, ranging from €7.5 million or 1.5% of global turnover to €35 million or 7% of global turnover, depending on the severity of the infringement and the size of the entity involved.

Implementation Timeline

The final text is expected to be published in the Official Journal of the European Union in early 2024, with a two-year period before it becomes applicable. Certain provisions will take effect sooner, including those concerning GPAIs, which will apply within 12 months.

Transparency obligations:?The AI Act imposes transparency obligations on providers and users of certain AI systems and GPAI models, including

1. providers of AI and GPAI systems generating synthetic audio, image, video, or text content,
2. deployers of emotion recognition or biometric categorization systems,
3. deployers of AI systems that generate or manipulate image, audio, or video content constituting a deepfake, and
4. deployers of systems that generate or manipulate text published with the purpose of informing the public on matters of public interest - content will have to be labelled in a machine-readable way so that it can be identified as artificially generated or manipulated.

The AI Act provides exceptions in some circumstances, including when the AI system is used for artistic, satirical, creative, or similar purposes.

Preparation for Businesses

Businesses are advised to start preparing for compliance by assessing their AI systems against the new regulations, developing AI governance strategies aligned with business objectives, and ensuring that their AI developments are compliant with both existing and forthcoming regulations. This includes evaluating risks, implementing robust policies, and considering both internal and external resources needed for effective AI governance.

This is a significant step that could set a precedent for AI regulation worldwide.

?

Challenges to Adoption

AI has been used in the BFSI sector for a long time, therefore the journey to Gen AI is relatively simple, however there are critical challenges.

·?????? Data Accuracy: when using LLMs that are trained on large volumes of data available in the public domain, the quality of the output is diminished and predictions are not accurate. Also the prevalence of bias in the data creates biased predictions. This will require continuous improvement.

·?????? Governance and compliance: BFSI is a highly regulated industry, and data movement must meet the rules and regulations and governance frameworks must be introduced and implemented.

·?????? Budget & justification: Deploying GenAI is expensive. Applying regulatory, compliance and governance framework is additional cost. The ROI is not clearly understood yet. It is clear that automation will impact functions and jobs, but whether that will generate net new revenue streams is not clear. Currently the prediction is that Gen/AI investments will reduce costs – which is different from prediction that it will increase revenue.

·?????? Talent shortage: This is a constraint across industries, not unique to BFSI.

·?????? Explainability – This is a very important factor. Explainability in the context of AI allows?human users to comprehend and trust the results and output created by machine learning?algorithms. Developing this capability requires understanding how the AI model operates and the types of data used to train it. Though that sounds simple and rather logical, the more sophisticated the system becomes, the harder it is to pinpoint exactly how it derived a particular insight. AI engines get “smarter” over time by continually ingesting data, both source data and user data, and fine-tuning and at incredible speed. So ‘explaining’ why we got a certain answer is not very straightforward. Further complicating the matter is that different users/use cases require different answers from potentially the same training data.

Ending this article here. Hope you learnt a little and enjoyed a lot. I did.

?

Additional reading

Generative AI in the Finance Function of the Future | BCG

MIT-UBS-generative-AI-report_FNL.pdf ( technologyreview.com )

The Impact of Artificial Intelligence on Jobs in the Financial Industry | LinkedIn

Role of #AI in #Financial Services and #Fintech | LinkedIn

?

Coming up Next

1. ?AI-griculture - the role of AI in Agriculture
2. AI in Cybersecurity | Auth0/Okta
3. The rise of Small Language Models
4. Where is Apple? Slow and steady wins the race?
5. The environmental cost of AI
6. Decoding ML, DL, LLM, AGI, and the world of AI
7. Rhythms | A new AI powered operating systems to transform the future of work?
8. The Chip War

?

?