Is Artificial Intelligence still relevant for Finance after Covid-19?
Digitization and adoption of AI had been very smooth till the pandemic began. Data needed for utilizing Machine Learning effectively was getting generated. But pandemic had a negative impact on Bank’s profitability, reducing the AI Budget Outlay. AI covers areas like Machine Learning, Deep Learning – Natural Language Processing, Computer Vision, Reinforcement Learning, Voice – Chatbot, Speech Recognition, Internet of Things, Robotic Process Automation etc.
Moreover, pandemic data being the noise or outlier in the training data, it is being discussed whether there is any value of the data gathered during the pandemic towards long term prediction.
On a positive note, from bank’s standpoint, during pandemic, banks used AI for fraud prevention and for advanced portfolio analysis. AI-based improvement in operational efficiency helped banks improve their service to clients. In addition, data generated during pandemic can help future decision-making to manage uncertainties in the environment.
Challenges of Finance Sector
- Managing environment uncertainties
- Exceeding expectation with unified customer experience
- Staying competitive
- Regulatory compliance
- Cybersecurity
To tackle these challenges, Banks and Insurance companies need to enhance operational and organizational efficiency. Finance companies really cannot escape passing through the stages of Digital transformation, Cloud migration, Automation, Data Analytics, AI/ML, Crypto/Blockchain adoption for coming to grips with the crippling challenges.
What are the main activities of Finance as an Industry Sector?
- Deposit/Lending
- Insurance
- Payment
- Investment and Wealth Management
- Capital Markets
- Trust, Transparency and Ethics
- Legal, Compliance and Risk Management
Commercial Banks and AI
Lending and Deposit
AI enables better prediction of customer behavior with using the following types of data along with AI/ML:
- Digital data – Web and Mobile data, ecommerce transactions
- Lifestyle data – Data regarding house, car and so on
- Geolocation and demographic data
- Psychometric data from Social Media – Attitude and Personality
- Financial transaction data/patterns – Cash flow data, spending patterns
Payment
Managing Infrastructure and Payment Channels – Commercial banks face huge challenge with two aspects of its functional need related to Payment.
- Fraudulent transactions – AI is very effective to detect and prevent fraudulent transactions.
- Network usage optimization – For example, AI identifies the pattern of ATM network usage, thereby can suggest the optimization regarding business process using the network and branch location.
Compliance/Regulation
Continuous monitoring of financial transaction and suggesting for audit creates a solid governance model.
Marketing and Sales
AI helps customer needs with granular details thereby making campaign/customer targeting more effective.
Personal Banking – AI based voice-activated interactive customer education/self-help and linking to related services based on customer’s identified need. AI-driven Personalized Wealth Management applications are gaining ground.
Robo-advisors/Chatbots are being used to provide portfolio management and wealth management services to customers. Robo-advisors consider customer behavioral data, investment objectives, risk profile to create optimal portfolio of the customer.
Role of FinTech
Image – FinTech Functions, Image Reference: The Impact of Artificial Intelligence on Governance, Economics and Finance, Volume-I by Sezer Bozku¸and Kahyao Glu
FinTech is an innovation for economy enables the use of technology for the Finance sector of the future. Fintech integrates technology into financial services. It helps technology adoption in Finance sector. The strong banking system, the expansion of card payment systems, and an intensive use of mobile technology provide important opportunities to FinTech.
“There are over 30 fields such as next-generation personal financial management, new digital lending, peer-to-peer lending and investment, mobile payments, aggregator comparison engine, mobile point-of-sale devices, international remittances, other payment processing, social integration IoT and connected devices, telematics, next-generation trade finance, prevention, next-generation collateral management, trading, trade analytics, peer-to-peer corporate lending and investment, one-stop shop for businesses, digital cash management, next-generation lending to SMEs, robo-advisory, crowdfunding, social investing, blockchain, investment across regions engine, payment infrastructure, big data base risk assessment, application programming interface ecosystem, anti-money laundering and know your customer, cybersecurity, artificial intelligence, and machine learning emerging as new norms in global banking and FinTechs provide services in these areas.”
Reference – The Impact of Artificial Intelligence on Governance, Economics and Finance, Volume-I by Sezer Bozku¸and Kahyao Glu
RegTech helps expedite the settlement of regulatory issues etc. InsurTech is helping Insurance sector in the same way as FinTech does for Finance sector.
AI and Credit Decisions
AI plays an important role in Credit authorization/Consumer credit decision, Bank loan losses prediction, Agriculture loan evaluation, Bond rating, Credit scoring, Sovereign credit rating, Credit risk classification in consumer loan, Corporate credit rating, Bankruptcy prediction etc. AI helps Mortgage credit decision-making and Real estate valuation.
Image – AI-enabled Realtime Credit Authorization Process, Image Reference – The AI Book by Susanne Chishti and others
AI and Fraud Detection & Prevention
AI has huge impact on controlling the rising credit card frauds of online transactions. The input data for AI-based Fraud detection system is clients’ behavior, location, computer IP address and so on. AI tools can offer market surveillance solutions to diagnose market anomalies/manipulations. It extends less-error-prone solution for know-your-customer (KYC) and anti-money laundering (AML) processes, It can identify inconsistencies and patterns that are highly complicated to spot manually. Further AI can offer quicker solution towards linking with ultimate beneficial owners (UBOs), politically exposed persons (PEPs) or prohibited states.
AI and Trading
AI monitors both structured and unstructured data (using Natural Language Processing – NLP) available to help the stock traders. Structured data here is the data captured with spreadsheet etc. that captures company’s financial details, and the unstructured data consists of the continuous information flow over internet/social media about companies. Nowadays AI is being used for Portfolio optimization/Portfolio management/Portfolio selection. AI can play important role in Asset Value prediction, Equity profitability forecasting, Earning Per Share forecasting, Index movement prediction, Consumer price index forecasting, corporate failure prediction, exchange rate forecasting, economic growth forecasting, index movement forecasting at the global level etc.
Algorithmic Trading – A prevailing trading mechanism for Institutional Investors. These tools quickly predict the stock prices for trading decision.
AI and Central Banking
Big Data-driven Analytics and Decision Support system is a boon to Central Bank’s policy makers. It enables supervision, surveillance, and monetary policy execution/effects on the financial system. AI is enabling better treasury management.
Risk Management
Risk Category
Market Risk – Risks/uncertainties of value of assets, liabilities, income etc. Includes interest rate risk, exchange rate risk, commodity price risk
Credit Risk – Also known and default and bankruptcy risk. Related to creditor’s ability to abide by financial contractual agreements.
Insurance and demographic risk – Related to Insurance industry sector. Variance of insurance claims. Also related to policy-holders demographic profiles like mortality etc.
Operational risk – Risk of loss due to unpredictability of business or loss dues to faulty business practices. Categorized in two types – Business risks and Event risk.
Image – Financial Risks and Machine Learning Algorithms manage those risks
ML Algorithms that take care of Functional Risks
Supervised Learning using Classification
- Fraud Detection
- Portfolio optimization
- Credit Scoring and Bankruptcy Prediction
Supervised Learning using Regression
- Volatility forecasting
- Sensitivity analysis
- Claims modeling
- Loss reserving
- Mortality modeling
Unsupervised learning using Clustering
- Insurance pricing
- Sensitivity analysis
- Credit scoring and Bankruptcy prediction
Anomaly detection
- Fraud Detection
Dimensionality Reduction
- Insurance underwriting
- Mortality modeling
Reinforcement Learning
- Portfolio optimization
Semi-supervised learning
- Sensitivity analysis
Reference – Machine Learning for Financial Risk Management: A Survey by AKIB MASHRUR , WEI LUO , (Member, IEEE), NAYYAR A. ZAIDI , AND ANTONIO ROBLES-KELLY , (Senior Member, IEEE)
AI and Insurance
AI can analyze/predict consumer behavior with effective product offerings.
- Underwriting – Quicker appraisal with less error-prone evaluation for effective underwriting.
- Pricing and Risk Management – Better Pricing and Risk management
- Claims – Quicker claim processing
- Better customer retention/marketing
- Operational efficiency and use of chatbots
- Cost efficiency and fraud detection for health insurance
InsurTech is working towards AI adoption in Insurance sector. Insurance sector is becoming so AI heavy that it is being called “Alogocratic” at times.
AI and Accounting
AI/ML is impacting the way account has been practiced. The main functional areas in Accounting with extensive use of AI/ML are as follows (along with AI-based tools that are being used):
- Account Reconciliation – Sigma IQ
- Account Receivable – HireRadius
- Citi’s Smart Match
- Account Payable – Medius, The Shelby Group, Basware, Esker, and AvidXchange.
- Audit – Effective use of Robotic Process Automation (RPA) in Auditing. Deloitte’s Argus.
Conclusion
Banks are going through AI-led financial sector transformation. In-app Banking Experience is changing. AI chatbots is being used extensively. Smart loan Applications are being utilized. But AI adoption for Finance towards an AI-embedded financial future needs a strong Governance and Policy Framework keeping the importance of the following in mind:
- AI Due Diligence
- AI explainability
- AI Biases/Ethical Challenges
- Algorithmic Collusion
- Systemic Risks and Data Risks
- Algocracy
- Cybersecurity
- AI Review Committees
- Information asymmetry and Digital financial inclusion