Banking payment data is the perfect arena for Artificial Intelligence (AI) to flourish. Characterised by a large number of transactions, related processes and customer interactions on a daily basis, payments generate huge volumes of structured and unstructured data. The perfect playground for AI.
AI can ensure customer payments are processed faster and friction-free (without any human intervention). Because they are Straight-Through-Processing along the payment journey of originating or intermediary and receiving bank, it enables customers to get their payment confirmation, statement information, and better liquidity information much more quickly.
Moreover, this helps banks reduce inquiry and investigation costs and helps customers manage their money faster and with more accurate information. Ultimately increasing customer satisfaction and retention.
But before you can use AI and NLP, data needs to be normalised and enriched. We talked about this in an earlier blog, but here’s a quick re-cap.
Normalisation is a technique often applied as part of data preparation for AI. Within payments this can be applied to normalise unstructured data such as party name and address into various ISO structured elements such as First Name, Last Name, Street Name, Postcode, Country etc.
Similarly, normalisation techniques can be applied to statements data and intelligent matching for faster updates to customers. What’s more, the same techniques can also be applied for payments enrichment to get correct bank codes such as BIC or country specific local RTGS or ACH clearing codes for faster payments.
Once your data is normalised and enriched it unlocks the vast potential of payments data.
Banks collect large amounts of data about their customers via various touchpoints. It is therefore vitally important that this data is collected, monitored, analysed and actioned upon. And since data comes in multiple ways and methods, many banks are finding it’s important to use AI to intelligently assimilate and use it for payments analysis and monitoring.
AI can significantly improve monitoring. From providing better customer service to categorising payments and statements through to intelligently matching them. Over a period of time, these monitoring activities will provide banks’ insight into customer behaviour with regards to their payment transactions in terms of seasonality, velocity, and peaks. It can also help banks and their customers to improve fraud detection and prevention. Banks can use payments data by country, currency, and correspondent to analyse customer payments behaviour – and in some cases even provide value added dashboards.
In addition, through valuable data insights, led by historical information and ISO 20022 payments data, a business can improve profitability, optimise revenue, and cut costs.
For example, banks can provide customers alerts on duplicate charges, unexpected increases in fees or delayed salary payments. You could even provide tips on how to reduce F/X rates, how to reduce fees or provide explanations on summary expenses by category. This turns banks from a commodity into a valued partner.
It makes banks highly proactive instead of reactive and even improves internal operations and processes. The possibilities are endless:
Lending
Automatic categorisation of various inflow and outflows
Analysis of income, expenses, and disposable income
Consistency and regularity of income, or missed income opportunities
Cashflow analysis
Trend analysis
Prediction of income and expenses
Cash balance predictions
Debt and savings recommendations
Based on the insights gathered for a customer and their peers, or based on the bank’s financial expertise, various recommendations can also be provided as very valuable guidance to the customers – especially to students, young professionals, and SMEs.
Automating an inefficient process simply leads to a faster inefficient process. That’s why AI needs to be deployed to ensure business processes are optimal. By using AI to perform continuous learning, machine readable data can then be used to update models dynamically.
This can support processes such as:
Operational efficiencies by reducing manual effort (e.g., auto repair and enrichment).
Reduced false positives while ensuring less friction for genuine customer transactions.
Creating personalised application processes based on historical data and risk assessment, helping to acquire more customers in real-time and without friction.
Essentially, this provides an increase in capacity without increasing manpower.
In addition to the obvious process optimisation gains, leading firms are now using various AI techniques such as machine learning and NLP to monetise the data itself.
AI has the ability to generate unique insights into customer behaviour for example related to:
Payments such as salary, vendor, F/X trade, remittances, trade, securities etc. allowing banks to monetise the data itself.
Banks can also compare various types of customers with peer group analysis to provide additional insight to customers using metadata.
For example, a corporate or SME sending various payments out at the end of the month that increases their need for enough balance on their account, while some of the payment receipts may not come until 5th or 7th of each month. This creates the need for an overdraft to cover the shortfall.
Using insight and analytics banks can offer new products to cover this shortfall, while extra cash in various accounts or currencies can be used to cover shortfalls or offer time deposits for additional interest. Using AI and data with insight means banks can offer targeted products to a corporate / SME for overdraft in the exact time and amount, or even offer a weekly or daily deposit product or smart F/X swap for multi-currency accounts.
The banks that get this right can tap into a range of benefits, from reaching previously unserved customers segments to opening up new revenue opportunities and even making product and market decisions driven by real-time usage analytics.
To unlock this value, partnerships with AI payments specialists are likely to be the way forward to capture that value.
Best of all, AI solutions can significantly enhance the customer experience. Payment transactions can include a lot of data, especially ISO 20022.
With the use of various AI techniques, banks can search rapidly and efficiently through this data beyond the standard set of factors like time, velocity, currency, beneficiary, types of transactions, country and amount. They can then analyse this data in new and innovative ways to vastly improve customer experiences.
As an example, transaction categorisation or tagging allows banks to understand customer behaviour such as what, where, when, and how and use information available in ISO transactions to offer new services and products. Using these AI techniques banks can help customers by being informative, providing guidance, and assisting them across their banking journey:
Inform
Provide tips on how to reduce F/X rate, how to do wire transfer, or reduce fees.
Alerts on duplicate charges, unexpected increase in fees or delayed salary payments.
Explanations such as a summary of expenses by various categories or breakdown of fees.
Guide
Provide suggestions on cashflow forecasts based on seasonal changes, overdraft protection on delayed incoming payments, or potential financial health checks for SME customers.
Assist
Assist in automation such as scheduling payments, sending payments reminders, or loan applications.
Most critically, systems can learn from each transaction, constantly improving and becoming more effective — something unique to machine learning and AI.
AI is undoubtedly part of the future of transforming payments. In fact, for many banks, ensuring adoption of AI technologies within payments is no longer a choice, but a strategic imperative. The huge volumes of data, interactions, processes, and transactions involved make it ideally suited for application of AI and NLP technologies.
With AI, banks will gain from huge data-processing capabilities at a low cost while customers will enjoy improved security and an enhanced customer experience. This also helps banks reduce the cost of manual intervention with the ability to increase capacity without any additional resources.
You’ll stand to make a good return on investment as well. According to analysts, by 2030, AI will save the banking industry more than $1trn.
From our experience, we know that the use of data and AI within payments can provide:
100% accuracy in quality of data
76% reduction in manual efforts
80% reduction in processing time
400% increase in capacity
40% customer satisfactions
26% improvement in margins
This is where Pelican AI’s solution solves and simplifies. It helps banks solve many of the issues related to payments processing.
It enables payments processing optimisation by performing automated and intelligent enrichment, routing, and repair. It also provides detailed analytics and a dashboard to gain insight into customers payments to offer better products and services.
Using normalisation, enrichment, categorisation, and recommendation, it provides better insight into customer behaviour. This means you can proactively inform, guide, or assist and in turn recommend products or services as a trusted partner.
Finally, for financial crime compliance Pelican AI’s solution helps to significantly reduce false positive rates and reduces manual intervention with less FTE while providing far more detailed analytics.
“We can finally see a future where the human factor is only involved at a supervisory level and all the data processing is done by machines. This removes bias, errors, and mistakes from our systems and speeds up the processes without increasing the risk of fraud.”
In a nutshell, AI is a game-changer. As technology gets more mainstream, I’m sure we will see even more innovative and creative uses of AI and NLP technology.
To learn more about how AI techniques in payment processing can reduce costs and improve the customer experience, speak to one of our experts today. Or, book a demo.