Home > What’s UP > blog > AI made easy: simple but rewarding AI use cases banks shouldn’t miss

blog

08. 09. 2018

AI made easy: simple but rewarding AI use cases banks shouldn’t miss

You’d be surprised how many easy-to-implement AI solutions are overlooked in digital sales despite the nice payoff they offer. Here is a basketful of low-hanging fruits in AI.

Companies tend to jump straight to the most challenging use cases of AI. Only a few of them consider applications that are not only easy to implement but also bring tangible results. Capgemini has found  that 58% of companies are tackling ‘need-to-do’ use cases, which are defined by high complexity and high benefit, but only about 20% are implementing low-complexity ‘must-do’ AI use cases, which offer considerable returns.

So what are the low-hanging fruits in AI?

Neglecting these ‘must-do’ AI initiatives can be a huge mistake, no matter the industry. But what are these applications anyway? Here are some examples:

  • Analysing consumer behaviour
  • Risk management
  • Reducing revenue churn
  • Forecasting
  • Contextual / predictive customer care
  • Facial recognition and consumer identification
  • Using a chatbot / virtual assistant
  • Product or services recommendation

Let’s delve into how these low-hanging fruits are picked in other industries! A global mining company, a client of Capgemini’s, used a low-complexity AI application to detect fault and measure performance. The company had detected quality issues too late during the manufacture of aluminium tanks before. By using an AI-based predictive model, they were able to optimise product quality, yield and energy consumption. Plus they can also predict product quality and product lifecycle with a 70% accuracy.

In another example, Swiss financial group UBS has implemented an automated trading program to deal with post-trade allocation requests from clients. The system scans client e-mails, looks for details on how they want to divide large block trades between funds, and processes and executes the transfers as well. For an investment banker, this would typically take about 45 minutes. For the app, it takes less than two. Meaning that bankers can spend their days doing more value-added activities, according to a report by The Financial Times.

The bump in new product sales can be nearly twice as big for high implementers of ‘must-do’ use cases than for low implementers.

Increasing focus on ‘must-do’ use cases improves benefits both in consumer-facing and operational activities. “There are a lot of benefits from AI; there is efficiency improvement, enhanced customer experience, speed to market. At the operational side there is optimization of operations, of workload, of credit card payment and issuance,” the head of data science at one of the biggest banks in Australia said in Capgemini’s survey.

Plenty of examples to learn from

There are myriads of ways for banks to leverage low-complexity and high-benefit AI use cases. Infosys Finacle reports that German digital-only challenger Fidor has just taken using customer behaviour analytics up a notch. They use data to create a community rating, named Community Karma, based on customers’ activities, connections and interactions. This helps them offer products that are truly relevant to their customers. Fidor Karma also creates a banking profile by integrating community contribution, social media profiles and connections with other community users.

AI can also help revitalize upselling and cross-selling, according to McKinsey. A top consumer bank in Asia boasted a large market share but fell behind in products per customer. They used advanced analytics to explore several sets of big data: customer demographics and key characteristics, products held, credit-card statements, transaction and point-of-sale data, online and mobile transfers and payments, as well as credit bureau data. The bank mapped out hidden similarities that allowed them to define 15,000 microsegments in their customer base. They then built a next-product-to-buy model that increased the likelihood to buy three times over.

India’s ICICI Bank has deployed software robotics in more than 500 business processes, covering a million banking transactions every day. Robotic process automation (RPA) has helped the bank cut the time needed to respond to customers by 60% and increased accuracy to 100%. ICICI Bank’s robots are working in a variety of retail banking operations, as well as in treasury and human resources management. They capture and interpret information, recognize patterns and run processes to perform functions like data entry and validation or automated formatting, Infosys Finacle reports.

Third-party providers to the rescue

Where do I sign up? Not so fast. First you have to find the AI applications that best suit your bank’s needs. Who can help? Some rely on their tried-and-tested in-house AI tools but more and more swear by teaming up with third-party providers. These companies can easily aggregate non-traditional data, such as customers’ online behaviour, social media interactions, geolocation data and even weather, with traditional customer information to provide better insights for sales.

Third-party insight-driven digital sales and engagement tools can help banks optimise sales opportunities and successfully compete with challenger banks and tech giants. W.UP’s Sales.UP application, for example, collects and decodes the digital signals banking customers send out every day. Based on these signals, banks can offer relevant products and services, having identified customers’ behavioural patterns or life events.

Pre-built insights in Sales.UP use AI and predictive analytics to show, for example, if a customer is planning to buy a house. How, you wonder? Algorithms not only analyse their account balance but also use non-traditional datasets like location, mobile device data and other data sources. And that’s not all. When mobile data is merged with transactional data, information like customers’ groceries spending, shopping habits or travels can be mapped out, and even visualised on a map.

ML tools need an enormous amount of data that is standardised, labelled and cleansed of anomalies. Getting internal data ready before moving into AI is a must. But there is help. There are more and more vendors taking public sources of data, organising it into data lakes and preparing it for AI to use, which might be yet another reason to opt for third-party providers. Sales.UP, for example, also helps banks with raw data collection, as well as data cleansing and aggregation.

For more on how financial institutions can benefit from AI tools download our white paper on AI in banking sales.

AI and ML in digital banking sales