“A couple of clicks, drag a slider here and there, a quick ID check, attach a few docs and bingo, the money arrives!”
– this is what users long for.
So what can a bank do to satisfy these desires while ensuring that its deposit account holders sleep soundly at night, too?
Winning over and retaining a digitally-literate generation has placed yet more pressure on banks. Full digitalization is unavoidable and not only because of external coercive factors: online credit scoring saves banks time and money,
the assessment processes are made more secure and the client circle ends up feeling more satisfied.
However, digitalization of credit scoring is not merely about transferring procedures that until now had been conducted in person onto a digital platform. The technology also promises new solutions that, in addition to identification and
administration, also position the determination of creditworthiness itself on a digital basis.
Data is the new oil, right?
Data analysis has become a global trend in all fields and a whole new profession has arisen in the shape of data analysts and data scientists. So why shouldn’t banks take advantage of this opportunity? Day after day, vast quantities of raw
data are generated worldwide, from which it is possible – using appropriate expertise and software developed for the purpose – to mine relevant correlations. Just imagine what the data registered by sensors in our telephones can reveal about us: daily commuting
patterns, preferred means of transport, destinations, demographic and health data, shopping habits. From a content viewpoint, our navigation through the internet is also very revealing, and not only our browsing history. There is a lot to be gleaned from,
for instance, how quickly we read, how we switch between sites and our data processing style. From the aspect of a data analyst and software developer, these are hugely valuable pieces of information.
It is absolutely inevitable that banks will also draw from this huge data pool, and if they could, they would certainly dip deep with the ladle. However, three factors are currently stopping them doing this:
- Consumer attitudes
- Lack of universality.
Why are utility bills and wage slips the documents that banks usually ask for?
Precisely for the three reasons detailed above:
- Because: This is the current legislation;
- Because: Consumers have no inhibitions about providing these data;
- Because: Almost everyone has these kinds of data and they supply an accurate picture of creditworthiness and standing in the community.
If any kind of change is implemented, it has to meet the above criteria. Still, from time-to-time, ApPello’s well established credit-scoring system is challenged, whether it knows already the innovations one could have read in Wired
magazine. Although we follow the trends, we are not willing to experiment for the sake of admiration. At heart, a bank provides a service of faith and trust, it bears responsibility for its customers, particularly as regards their data. Therefore,
it is critical to consider the limit where the use of data improves the quality of services provided by banks without too much personal data leading to mistaken conclusions. If a bank involves data lifted from social media into the risk assessment, then it
has to have a coherent scoring system. It is obvious that this evaluation process still requires refinement of the models and optimization of protocols. Unlike Big Tech enterprises (Apple, Google, Facebook etc.), banks are not yet in possession of
such giant data troves. However, if regulations were modified to allow the above mentioned Big Techs to function as financial institutions, it could result in a totally new competitive situation. They would instantly have a huge advantage particularly
in the retail sector as they would be able to provide a much more customized user experience based on the richness of the profile.
The next unavoidable issue is consideration of the rights of the individual. Customers should also weigh up the extent to which they provide authorization for the handling of their own data if this eases or speeds up, for example, the assessment of credit
applications. The ‘millennial’ generation, i.e. digital natives, are not so reluctant because they have socialized in a data-driven society where exploiting the benefits inherent in this takes precedence over privacy protection.
Another fascinating topic on the subject of digitalization of credit scoring is the application of artificial intelligence. It would seem obvious to involve AI in data analysis during credit assessment, but there are concerns even on the
side of regulators, and not without justification. If banks use a model based on a neural network, it is difficult to point out reasons for any given decision at a later date. This is because AI presupposes a machine learning
process in which the system learns to draw the appropriate conclusion from a relevant large sample, but it is precisely for this reason that it will never be able to pinpoint with any accuracy the underlying reason. Obviously, this can result in legal
consequences since a customer never receives an objective answer as to why his/her request for a loan was turned down. Every financial institution needs to be particularly prudent because managing the money of customers is a matter
of trust that is easy to lose even over a minor mistake.
However, there is no stopping technological development.
Today, the banking sector employs solutions on a daily basis that were not available five years ago, and ten years ago were simply inconceivable. In addition to thoroughly reworking the regulatory background, another challenge in digitalizing the
financial sector is to attract trained IT professionals. At the moment this presents the banking sector with a massive HR problem because it is not necessarily the financial sector that offers the most exciting career pathways for the best IT experts
and it is far from being the best paying field. On the other hand, as the attitude of banks changes and they become ever more committed to full digitalization, they will inevitably devote greater resources to data processing. The longer a bank puts
off learning about what is new in this field, the higher the cost will eventually be.