Keeping An “AI” On Fintech: AI-Based Use Cases Poised To Take Financial Services To The Next Level
To get a picture of the momentum behind AI, the global artificial intelligence market was valued at $62.35 billion in 2020 and is expected to expand at a compound annual growth rate (CAGR) of 40.2% from 2021 to 2028. Given this projection, it’s not surprising that tech giants such as AWS, IBM, Google and Qualcomm have all made significant investments into AI research, development, disparate impact testing and auditing.
My coverage area of expertise, fintech (financial technology), is no exception to this trend. The AI market for fintech alone is valued at an estimated $8 Billion and is projected to reach upwards of $27 Billion in the next five years. AI and machine learning (ML) have penetrated almost every facet of the space, from customer-facing functions to back-end processes. Let’s take a closer look at these changing dynamics.
What AI and ML use cases are applicable to fintech?
Financial services have their own set of common AI and ML use cases. These include, but are not limited to cost reduction, process automation, spend reconciliation, data analysis and improving customer experiences. According to a Bain report, the pandemic has widened the gap between top performers and other companies in three main productivity drivers—people’s time, talent and energy. These AI use cases stand to help address the business challenges associated with this gap.
Process automation, one of the industry’s most common use cases, reduces the amount of manual work done by employees by automating repetitive tasks and processes. Not only does this reduce tedium, but it also gives workers time and energy to focus on innovation and other value-adding activities, which can also boost morale.
The processes of spend reconciliation and payment authorization are also highly labor intensive and time consuming for accounting departments. AI and ML can provide automated, three-way matching of incoming invoices from suppliers for approval. Intelligent systems can also learn the approval process for more complex and fragmented areas of expense spend, such as the miscellaneous travel, goods, and service expenses of today’s increasingly distributed workforce.
For example, Beanworks, an accounts payable (AP) automation provider, recently introduced SmartCapture, an AI-driven data capture functionality that promises to drastically increase customers’ data entry speed and accuracy. The company claims its offering provides over 99% accuracy while completing AP processes in minutes. Combined with its SmartCoding technology, it claims to reduce the time accounting teams spend on data entry by more than 80%.
According to Beanworks, traditional, manual AP data entry is inefficient, error-prone and consumes considerable time and resources, costing anywhere from $12-20 per invoice to process. An AI solution such as SmartCapture, according to Beanworks President and COO Karim Ben-Jaafar, gets smarter with every invoice as it learns how to correctly interpret an organization’s accounts payable documents and coding. This approach, said Ben-Jaafar, “frees up accountants’ time and energy a to focus on other, more strategic tasks.”
Other applications of AI and ML help consumers better manage their finances with hyper-tailored offers that fit their current financial situation, including credit on-demand and at a lower cost.
In a recent column on the pandemic’s impact on credit trends, I quickly touched on Upstart, an AI-lending platform that is reimagining creditworthiness. It’s worth taking a deeper look at the company’s efforts to overhaul our outdated credit scoring system, which, as it stands, makes it difficult for lenders to accurately assess which borrowers are likely to default. This uncertainty often results in consumers being denied access or overcharged for credit. Meanwhile, Upstart’s system leverages AI to approve more than two-thirds of its loans instantly.
Engaging AI credit-scoring software can reduce non-performing loans while boosting returns , meaning better loan decisioning. This means companies lend to less risky clients, and customers can access personalized, almost instant loan decision-making.
In keeping with the finance on-demand theme, two of the most sought-after AI/ML solutions are personalized portfolio management and product recommendations. Their popularity is growing, as is their refinement. Investment platforms such as Betterment recommend clients' investment opportunities based on income, current investment habits, risk appetite and more. The wine investment platform, Vinovest, uses custom machine learning algorithms developed with world-class sommeliers to curate a wine portfolio that acts as both an alternative investment that can out-perform other asset classes and a “liquid asset” should the user actually decide to drink their wine.
In the coming years, we are likely to see refinement in all of these technologies. I believe we’ll also see increasing amounts of automation in customer support, report generation and data analytics.
Can “it” be trusted?
As financial services providers continue on their digital transformation, we will see an increase in AI and ML security solutions over the next year. For example, I expect to see more regulatory (RegTech) solutions that analyze documents for account registration, detect anomalies in patterns within accounts and more. AI and ML moved from a curiosity to a necessity and priority during the pandemic, particularly for financial services.
AI needs auditing—we can’t assume what the machine is learning will always be correct. As long as the proper audits are in place, though, AI can bring security to financial services. Granted, any technology used in modern banking demands a strong regulatory lens. Creators of AI in financial services have to prioritize the audit trail while constructing their solutions.
Gene Ludwig, founder and former CEO of Promontory Financial Group, an IBM-owned financial services consulting company, summarized this well during a Reinventing Financial Services Digital Forum podcast. “AI, properly documented, is actually better than the human brain because you can’t get inside the human brain and document the decisioning pattern that’s used. It’s actually potentially much better, not worse than people in terms of audit trails.”
A gratifying result
What we’ve learned from the rise of fintech apps is that customers crave instant gratification—whether in response time or personalization of their experience. Refinement in AI and ML will give the leaders in the space those precious minutes or even seconds that it takes to compete for someone’s business while delivering a personalized product to them. From customer-facing functions to back-end processes, the possibilities are endless.
Disclosure: My firm, Moor Insights & Strategy, like all research and analyst firms, provides, or has provided research, analysis, advising, and/or consulting to many high-tech companies and consortia in the industry. I do not hold any equity positions with any companies or organizations cited in this column.
Post a Comment