Mahi Sall, Advisor, Fintech-Bank Partnerships, Payments and Financial Inclusivity
January 25th, 2023
Mahi Sall is the CEO & Founder of FinXpair, a Berlin-based strategic advisory boutique firm focu[...]
Classic Cash Forecasting vs Powering with AI
Guest Post | Jan 31, 2021
Financial institutions that deploy ATMs face numerous challenges. Even in this day and age, when every field is undergoing digital transformation and people are carrying out their tasks online, installing and managing ATMs continue to be imperative to banks. During different kinds of crises such as the 2020 global pandemic, ATMs served as substitutes for bank branches as the offices were either closed or had shorter working hours. These locations provided customers with numerous services they could avail without encountering the risks associated with human interactions.
One of the biggest goals to keep up when managing ATMs is ensuring customer satisfaction throughout the week. Banks do not want to let their customers return empty-handed. But it is also difficult to tread the fine line between cutting down costs and ensuring consumer satisfaction.
Cash forecasting in the past
Many ATMs often run out of cash much quicker than anticipated, leaving loyal customers no other alternative but to visit ATMs belonging to other banks. Some ATMs in specific regions are less frequented by users, in comparison with the others. Banks often forget to distinguish between the two and fill all different kinds of ATMs with the same amount. This leads to excess amounts of idle cash lying unused in specific machines. The cash, if it was not deposited in the machine, could have easily been used for profit-generating initiatives such as loans.
These issues made it essential for banks to have an effective cash management system in place that could optimize cash in all their ATMs. To do this, many financial institutions partnered with atm services that offered banks numerous advantages.
In the past, the employees at the institution manually determined cash requirements at ATMs. They based these off on general trends that could be easily deciphered such as increased number of withdrawals during the holiday seasons, ATMs near shopping malls witnessing more users during the weekends than the weekdays, etc. The method was never free of errors, and banks often required their CITs to refill machines that had unexpectedly turned empty. Many also tried applying the pattern-based replenishment schedules to several machines at once. This certainly did not work as withdrawals at each ATM varied according to the ATM’s location, the time of the day and the background of those who usually visited the premises.
Technology turned out to be a fantastic boon for banks struggling with managing the cash levels in their ATMs.
Traditional cash forecasting
The replacement of human cash forecasting with algorithms and machine learning began towards the end of the 20th century. A few companies took the initiative to fine-tune the cash forecasting process by removing human error and using Artificial Neural Networks (ANNs). These algorithms that are commonly used in various forecasts in the financial industry considered multiple factors before predicting accurate cash requirement levels with minimal margins of error. A skilled data scientist was employed to configure these systems for developing cash forecasts regarding each machine. Banks no longer had to group together machines based on common assumptions that lead to higher costs.
See: How AI is impacting customer experience in fintech
The software analyzed the historical withdrawal data of every machine to understand how withdrawal trends changed daily, weekly, monthly or during different seasons. With insights from the system, banks could now fill each of their machines with the exact amount it required and not worry about running out of cash unexpectedly. CIT schedules were revised to make sure that ATMs were refilled just before they ran out, and adjusting the money filled in each machine meant that the schedule could be improvised to enable the filling of multiple machines on the same day.
In addition to not encountering emergency supply runs, banks also no longer had to worry about having excess funds sitting around in their machines. All the cash that was not being utilized in ATMs was now used in their proper channels to bring in profits.
AI-powered cash forecasts
One of the biggest drawbacks to traditional forecasting systems was that they used historical data to predict future requirements. When the 2020 pandemic emerged and shook the world, ATMs saw extremely volatile and unfamiliar withdrawal patterns, unlike what was followed previously. The forecast systems no longer had months of historical data to create accurate predictions. Withdrawal trends changed every day, and due to widespread panic, many started withdrawing considerably larger amounts. Banks had to divert money meant for less used ATMs to machines that were utilized by many.
Read: FFCON20 Week 7 Wrap-up: Artificial Intelligence in Fintech
This period of volatility led to the necessity of more advanced forecast systems that were equipped to deal with unexpected changes. Artificial Intelligence-based systems are different from the traditional ones and make use of day-to-day withdrawal data to come up with accurate predictions. The latest systems use a nonlinear methodology that can identify and adapt to unexpected withdrawal demands. The cash demand data produced by many institutions contain numerous flaws, but the forecast system comb through these to remove the abnormalities and come up with correct values. They can also take into account sudden changes in the environment that could shift withdrawal patterns to either extreme.
Summing up
Banks have come a long way from using human minds to calculate ATM cash requirements. Today, most banks use highly advanced systems that analyze historical data to develop accurate cash predictions for individual ATMs. But recent events have led to the need for a better system that analyzes real-time data to predict future withdrawal patterns. Software systems that use artificial intelligence are proving to be indispensable in this area.
The National Crowdfunding & Fintech Association (NCFA Canada) is a financial innovation ecosystem that provides education, market intelligence, industry stewardship, networking and funding opportunities and services to thousands of community members and works closely with industry, government, partners and affiliates to create a vibrant and innovative fintech and funding industry in Canada. Decentralized and distributed, NCFA is engaged with global stakeholders and helps incubate projects and investment in fintech, alternative finance, crowdfunding, peer-to-peer finance, payments, digital assets and tokens, blockchain, cryptocurrency, regtech, and insurtech sectors. Join Canada's Fintech & Funding Community today FREE! Or become a contributing member and get perks. For more information, please visit: www.ncfacanada.org
 |  |  |
Support NCFA by Following us on Twitter!Follow @NCFACanada  |
Related Posts
Tink Survey Report 2021: Open banking spending, investments and use cases- Reg CF: Investment Crowdfunding Tops $1 Billion on Heals of Funding Cap Increase
- Ways you can help support Ukraine
- The Good, the Bad and the Ugly of Central Bank Digital Coins (CBDCs)
- The AI Revolution in Wealth Management: Top 3 Innovations (and more)
- IIROC Grants Fidelity Crypto Asset Trading Insurance and Custody Exemptions
Leave a Reply