About the author: Adhikar Babu is a Senior Product Manager at Worldpay with extensive experience from McKinsey in strategy and product roles. He has successfully developed products in industries ranging from Automotive SaaS to financial services, driving a revenue increase of over USD 70 million across multiple product lines. Adhikar’s leadership and innovative approach significantly impact the tech and financial sectors.
As we can see already now, AI is set to revolutionise various aspects of daily life, including our purchasing habits and payment methods. Artificial intelligence presents evident opportunities for consumer-facing operations, such as online payment methods, and significant potential for enhancing operational efficiency within fintech companies. The AI fintech market was valued at $6.67 billion in 2019 and is expected to reach $22.6 billion by 2025, with a compound annual growth rate (CAGR) of 23.7%.
After the outbreak of the COVID-19 pandemic, we witnessed a major change in the way people engage with financial services and in their financial behaviour in general. This was influenced by various reasons: a reduction in cash inflow due to income shocks, increased cash outflow from pandemic-related expenses and additional responsibilities, and wealth reduction resulting from depleted savings and declining investment values.
Additionally, the pandemic has accelerated the digitisation of economic activities, such as shopping and banking, as social contact has been minimised, further affecting daily financial management practices.
It is not surprising that fintech firms specialising in payments and wealth management have prioritised fortifying their infrastructure, investing in new resources, and expanding capacity to accommodate increased transaction volumes.
While these issues posed challenges, they highlighted the critical role of AI solutions, vital for fintech companies reliant on transaction volumes for revenue. These developments are poised to drive the demand for AI solutions in the fintech sector. Below, you will find an overview of the opportunities that have the best prospects and the challenges we may face implementing them.
How AI Affects the Industry
AI has a significant impact across all facets of the fintech value chain. Below are the primary use cases illustrating the potential for “AI-fication” in various areas.
Enhancing Sales
AI facilitates assisted recommendations during client interactions, personalised product offerings, and dynamic pricing, potentially increasing sales rep productivity by 28% to 38%. Key account managers could see even greater productivity gains through advanced automation of processes like request for proposal (RFP) creation.
Improving Risk Management and Compliance
AI aids in fraud detection and prevention, streamlines monitoring and reporting, and enhances trend prediction and anomaly detection. This results in cost savings, error reduction, and better resource allocation.
Efficient Product Development
AI significantly impacts coding excellence, improving productivity by 28% to 38%. It enhances coding quality controls, reduces technical debt, and assists both junior and senior developers in various tasks, although scaling AI-supported coding across the organisation presents a challenge.
Tailored Solutions for Individual Needs
With the help of AI capabilities, fintech organisations can anticipate and cater to the unique needs and preferences of customers, offering personalised recommendations and value-added services promptly. This customised approach not only enhances engagement but also stimulates sales growth.
Advanced Security
Fraudulent activities like credit card and payment fraud pose significant threats to fintech businesses, with over 70% reporting losses totalling approximately $500 million in 2022. As attackers increasingly employ sophisticated tactics and exploit emerging technologies, traditional security measures fall short in several key areas and AI offers solutions to these challenges:
– Advanced Behavior Analysis
AI analyses historical behaviour data to detect anomalies and potential security breaches early on.
– Encryption Standard
AI updates encryption protocols to improve platform security and reduce the risk of data breaches.
– Compliance Check
It quickly assesses regulatory updates and compliance status, ensuring adherence to evolving regulations.
– Real-time Transaction Monitoring
AI identifies fraudulent patterns and triggers immediate preventive actions to block fraudulent transactions in real time.
– Third-party Risk Assessment
Artificial intelligence apps evaluate third-party systems’ financial stability, cybersecurity posture, and regulatory compliance to eliminate vulnerabilities.
Moreover, AI technology in finance reduces false positives, identifies previously unseen threats, and offers substantial cost savings. For instance, implementing tools like Mastercard’s Consumer Fraud Risk tool could collectively save UK banks around $125 million per year on scam payments.
Resource Optimisation
AI technology also influences finance workflows. Fintech companies eradicate manual tasks, streamline business processes, alleviate talent shortages, identify errors, and optimise human effort. This allows employees to concentrate on high-value tasks requiring human input, while mundane, time-consuming activities are delegated to AI systems.
Using Data for Informed Decision-Making
In the finance sector information accumulates rapidly. With AI’s capacity to meticulously analyse vast datasets, financial institutions gain comprehensive insights into various facets of their operations. This helps them to formulate strategies and drive data-centric decisions. Indeed, with the high stakes in the business landscape — with top companies squandering approximately $250 million annually due to subpar decision-making — there’s an increased urgency to augment human expertise with the precision, swiftness, and accuracy afforded by AI-driven data analysis.
Limits of AI in Fintech
Despite its benefits, AI in fintech also presents bias concerns and the need for regulatory changes.
1. Biases in AI System
AI systems may inadvertently perpetuate biases in the data used to train them. As a result, certain minority or marginalised groups may be excluded or disadvantaged when accessing fintech services such as eKYC (electronic Know Your Customer) or facial recognition.
This exclusion occurs due to inadequate representation of these groups in the training data and the inherent biases encoded within it. Suppose the historical data used to train the AI model reflects existing biases in the lending industry, where certain demographic groups have historically faced discrimination or systemic barriers to accessing credit. As a result, the AI model may learn and perpetuate these biases, leading to discriminatory outcomes in credit decisions.
Moreover, since AI is trained on internet data, it suffers from “hallucinations” (making up random statements often stated in a matter of factly way) – fintech applications demand 100% certainty and a high degree of accuracy is it is a highly regulated industry
2. Superficial Mimicry of Human Thought
Another limitation of AI in fintech is its tendency to mimic human thought superficially, without possessing a deep understanding or innovative problem-solving capabilities. While AI algorithms can process large volumes of data and perform complex tasks, they often lack the nuanced understanding and creative problem-solving abilities inherent in human cognition. This limitation can hinder AI’s ability to adapt to new and unexpected situations or make contextually appropriate decisions, particularly in complex financial scenarios.
Another issue here is that AI models are still not exposed to financial data, which is typically hard to use. Financial data sometimes simply does not provide proper context for AI to come to any reliable conclusion.
3. Regulatory Challenges
AI analytics in fintech may also present regulatory challenges due to a lack of transparency in their decision-making processes. The problem is that AI algorithms often operate as black boxes, making it difficult to understand the reasoning behind their decisions.
Preventing Future Implications
Achieving broad adoption of Artificial Intelligence in fintech requires several critical improvements:
1. Standardisation efforts within the industry to ensure structurally similar data is available across multiple payments companies.
2. Training AI models on broader and “clean” financial data for more effective supervised learning.
While this can seem easier said than done, breaking down the problem into manageable steps could help. For instance, open banking initiatives could prioritise enabling consumer-level data access initially, while fintech players work on upgrading their infrastructures to unlock industry-level data.
Prospects
Several strategies have already demonstrated success in leveraging AI within fintech. In product development, payment companies have introduced “AI Copilots,” such as Stripe’s AI-powered developer documentation, facilitating coders with queries like API usage and sample code requests. Additionally, in transverse activities, firms have implemented AI chatbots like Klarna’s, claiming to perform the workload equivalent to hundreds of customer support agents. Moreover, in risk and compliance, AI, particularly machine learning, is widely employed for transaction monitoring and fraud detection, effectively identifying patterns and flagging suspicious activities.
Today many funds focus on seed-stage companies democratising financial services through AI. In their turn, numerous startups have already taken this advantage and leveraged AI to innovate in the finance sector. Below are just a few examples.
Stripe: Introduced Stripe Tax, leveraging AI to simplify tax calculations and compliance for CFOs and businesses operating internationally.
HighRadius: Created LiveCube, a no-code platform for CFOs to develop applications independently, reducing reliance on IT departments.
Paro: Raised $25M to expand its platform connecting businesses with freelance accountants, streamlining financial talent acquisition.
DataRobot: Specialises in customer needs-focused AI tools for retail and healthcare, aiming for broader industry applications.
Datarails: Launched “Genius,” an AI-powered financial planning and analysis tool, offering insights and assistance to finance professionals.
Ramp: Acquired Cohere.io to enhance finance automation with AI, aiming to unlock insights and improve customer experiences.
AI in Fintech Case Study
Moreover, many prominent companies have already employed AI and are successfully using it to solve repetitive tasks. They use artificial intelligence to automate documentation management, for instance, compiling user-friendly documents like terms of service and privacy policies. Besides, AI enhances Know Your Customer processes by automating identity verification and compliance checks. AI-powered chatbots provide real-time customer support and help to reduce the burden on the support service. Banks tend to use AI financial advisors to offer personalised investment recommendations based on individual preferences and financial goals.
Klarna AI Customer Operation Agent
Let us take a look at Klarna’s AI Customer Operation Agents. Klarna is a Swedish fintech company that provides payment processing services for the e-commerce industry, managing store claims and customer payments. Besides, the company is a “buy now, pay later” service provider.
In February 2024 Klarna announced its AI assistant powered by OpenAI. Now when it has been live for quite a while, it has become obvious that it is revolutionising customer service by handling two-thirds of customer chats in its first month.
With 2.3 million conversations and the equivalent workload of 700 full-time agents, it boasts impressive efficiency. It matches human agents in customer satisfaction while outperforming in errand resolution, leading to a 25% drop in repeat inquiries. Customers now resolve issues in under 2 minutes compared to the previous 11 minutes.
Available in 23 markets, 24/7 and over 35 languages, it’s estimated to drive a $40 million profit improvement in 2024. Moreover, it enhances communication with local immigrant and expat communities.
Klarna’s CEO emphasised the significant leap forward in the vision of a fully AI-powered financial assistant, aiming to save time, worry, and money for consumers and boost industry efficiency.
Stripe’s AI-Powered Developer Documentation
Stripe has introduced an innovative feature to its documentation, it allows developers to interact with it using natural language queries.
Through GPT-4 integration, developers can ask questions like “What is test mode?” or “How do I test my Stripe integration?” and receive concise summaries or specific information extracted from the documentation. This enhancement aims to improve the developer experience, enabling them to spend less time reading and more time building.
<iframe width=”560″ height=”315″ src=”https://www.youtube.com/embed/nx34XNBKno0?si=8ESwWj2H9TeeAXMG” title=”YouTube video player” frameborder=”0″ allow=”accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share” referrerpolicy=”strict-origin-when-cross-origin” allowfullscreen></iframe>
These developments reflect the increasing integration of AI in fintech, focusing on enhancing efficiency, accessibility, and strategic decision-making in the sector. Yet, while numerous fintech companies heavily invest in AI technologies, consumers often lack awareness of these advancements. Hence, there is a critical need for effective communication between businesses and their customers to bridge this gap in understanding.
To Conclude
Generative AI possesses the potential to propel financial innovation forward, a prospect both exciting and alarming. While the advancement is promising, there’s apprehension about how AI processes financial data, given its profound implications for end users. Imagine an AI-powered loan office deciding the fate of your home mortgage application.
Though we’ve only begun tapping into AI’s capabilities for financial use, progress is underway. Stay tuned for further developments in this evolving narrative.
