The role of AI in spurring financial inclusion

The financial services community is abuzz with chatter about how artificial intelligence (AI) is transforming banking, investment management and insurance. Yet much of this tends to focus on developments in advanced nations, be they in the West or in Asian economies such as mainland China, Hong Kong or Singapore.

A core benefit of AI to financial institutions is its ability to lower operating costs. Combined with robotic process automation (RPA), AI can bring about once unthinkable efficiencies and automated tasks. Hence, the technology can play a significant role in realizing financial inclusion among emerging markets.

Take customer interaction, for instance. The deployment of virtual customer assistants offers an alternative to information desks, tellers, financial advisers and call center staff. Using such tools, which are actioned through text or speech, enquiries are processed quickly and effectively, connecting customers to appropriate services while cross-selling additional products.

In South Africa, machine learning specialist DataProphet is working with local insurance companies to manage customer enquiries with chatbots. Having identified customer support as the second-largest cost of serving customers, the firm is developing a platform-agnostic solution that can be accessed via text, online or mobile, for a fraction of the cost that such services usually demand. This is potentially a game changer for swathes of the population who cannot afford conventionally priced insurance products.

Aside from pricing, there are other benefits. Customers who are less literate, or have sight or hearing difficulties, can leverage voice recognition to communicate their intentions. In India, for instance, 28% of Google searches are now conducted by voice. This demonstrates the potential of these technologies to financial providers.

With products at significantly lower price points in such markets, providers can leverage AI-powered tools to increase personalization and cross-sell new and cheaper services. In India, ICICI Bank has created iPal, an AI-based virtual chat assistant. Launched in February 2018, iPal has interacted with more than three million customers, and answered about six million enquiries with a 90% accuracy rate. The service assists customers in three areas: answering common questions about their accounts and other services provided by the bank; executing transactions such as bill payments and fund transfers; and offers guidance with tasks such as activating credit cards and resetting online banking passwords. The bank says it is working with existing voice assistants such as Cortana, Siri and Google Assistant to broaden the service and make banking accessible to all.

Arguably, the most significant advancement AI brings to financial services providers is its ability to eliminate against biases. In risk profiling, banks and insurers currently use a fixed number of predetermined controls to assess the credit worthiness of a customer, and use these to instantaneously decide whether an applicant is suitable for a particular service. AI has the ability to source information that is far beyond what conventional checks can find, and can examine other information, which gives a more informed view of the customer’s credit profile.

In Kenya, for instance, mobile financial services providers Branch and Tala leverage data collected by smartphones — including call log, demographic details, airtime purchases, financial transactions, and social media activity, among other information — to determine a customer’s creditworthiness and dispatch a microloan within minutes. Interestingly, the same methodologies are also being applied to detect fraud and credit anomalies. 

Similarly, with banks and insurers basing their product suitability assessments on items such as income and spending habits, they are able to target customers who are most likely to use a particular service. Yet by exploring other information and viewing customers in a new and unbiased manner, they can bring products to customers who would otherwise be excluded, and boost their profitability at the same time.

While the majority of AI and associated technology discussions focuses on opportunities, risks are nonetheless prevalent. These, if not managed well, could result in corporate failures and social harms. Three considerations are noteworthy:

1. AI, like all technologies, carries a risk of performance failure, and spotting imperfections within the technology’s myriad complexities is not easy
2. AI is only as good as the quality of data that is fed into any such system and the people who program them, which can easily discriminate or favor biases
3. Like any tech solution, cybersecurity and data protection is paramount to defend against malicious actors and prevent legal and reputational challenges

To manage these risks, financial providers and their technology partners must ensure that there is a governance framework in place, detailing all accountabilities. Furthermore, an extensive quality assurance program must be drawn up to ensure that any use of AI has clear objectives, is secure from all risks, and is compliant with local laws.

Providers must also engage with regulators. This will not only ensure that they remain on the right side of the law, but it is also an opportunity to educate authorities on how AI works in practice, and address imperfections that might prevent the technology from advancing further.

For now, however, AI is staking a credible claim in furthering financial inclusion. With the support of the entire financial community, its deployment within various products and services across emerging markets looks set to grow exponentially in the near- and long-term future.

You can the latest on financial inclusion at ey.com/financialinclusion or talk to me about how EY can help your business.

The views reflected in this article are the views of the author and do not necessarily reflect the views of the global EY organization or its member firms.