Generative AI Explainer
As indicated by the name, this is Artificial Intelligence that generates. What GenAI produces is data, in the form of texts, images, audio, video etc. Examples include concepts like ChatGPT, and what anyone with a social media account cannot escape, Midjourney, which generates hyper-realistic imagery. This is probably the part of AI generating the most attention as the output is human-like creations like «art» and coherent text.
The fact that AI can, not only mimic human creation and logic but go beyond this level to develop new processes, products, and levels of productivity is on the verge of paving the way for fundamental change across almost any sector of science, business, education etc.
Fraud prevention
False declines, companies lose approx 3% of their revenue yearly due to false card declines, legitimate transactions flagged as fraud. AI-based algorithms can correctly identify transaction anomalies by taking advantage of Machine Learning instead, rather of a rule-based, algorithm that has a limited capacity for fraud detection. Using AI can enhance accuracy, increase efficiency, reduce cost, and facilitate real-time monitoring.
Global digital fraud losses are projected to exceed $343 billion between 2023 and 2027. (source: Juniper Research) It is no wonder that the financial industry has a lot to gain by taking advantage of AI in the context of fraud prevention! In addition to using Machine Learning to detect anomalies, one can use Conversational AI for voice authentication through voice biometrics.
A possibility that leads us to the next steps of having a Know Your Customer (KYC) process that is watertight.
Finance Institution Regulation
Know Your Customer and Anti-Money Laundering (AML) regulations represent substantial costs for most companies in the payments industry. Not following said regulations could lead to astronomical costs in the form of fines or even the termination of authorisations to operate in the sector.
Machine learning has great potential to lead companies through the labyrinths of regulatory compliance. As machine learning technologies «ingest» new regulations, ML will be able to identify where it’s applicable and gaps in the existing regulations program.
However, since AI is intrusive by nature, especially AI based on machine learning, it raises complex issues, particularly with regard to the protection of personal data, and AI players need to pay particular attention to the legal issues surrounding personal data.
Care must be taken not to infringe on the rights and freedoms of individuals when designing and using these tools.
The approach adopted must take account of the fundamental principles of the GDPR, in particular the lawfulness, fairness and proportionality of processing.
Code Generation for Payments Products
Succeeding in Fintech means succeeding with IT development. This, obviously, comes with a price. A price that AI code generation can bring down drastically.
Probably 99.9% of data programming is based on already established methods and procedures. This means that the data is «there for the taking», and adding machine learning on top means that there is limitless potential for developing systems, programs, and products for payment solutions.
There is already a wide range of AI code generators on the market, including one from OpenAI, «OpenAi Codex». When using these kinds of tools, as with any other types of data, one must be cautious regarding what the input is to the tool and ensure that you are not sharing confidential company data.
Low code/no code AI refers to tools that allow anyone to create applications with virtually no coding involved. One example includes Apple CreateML which allows the user to create iOS applications in a simple drag-and-drop interface. Google and Microsoft offer similar solutions, and a tool like PyCaret lets the more advanced user create in the popular Python programming language with pre-configured functions.
In short, AI and ML are completely changing the landscape for creating applications and programs, across the board, and will have a profound impact on payments as in any other tech-driven industry.
Support & Onboarding
Chatbots, remember when they were only dumb and annoying? Well, this is changing. Applying machine learning means that chatbots are able to handle increasingly more complex queries from users. This means that more and more support issues can be handled «at the door», and do not need to be passed on to support professionals. In fact, the speed and availability of bots should also improve the user experience!
Onboarding is another resource-demanding practice that can benefit greatly from automation through AI, for instance, to automate the verification and authentication of customer data. Chatbots or AI-generated “hosts” can guide customers through the process. The human “touch” is only necessary for the final quality control.
It is fair to say that the more data your company is handling, the more machine-learning solutions will become central to your business. In the world of payments, this is perhaps as challenging as it is necessary due to the restrictions placed on data usage. What you cannot afford is not to have a strategy regarding artificial intelligence.
