CEE21: Why we shouldn’t be afraid of using AI?

During the latest CEE21 SME Banking Club Conference, Patrycja Sobczyk (Product Owner at Asseco Poland) talked about the usage of artificial intelligence in financial institutions.

This year, Asseco Poland and SME Banking Club produced a joint Study on the topic of AI in the SME Financial Sector (you can download it for free here), which showed that only 1/3 of the SME financial institutions use AI. 

What are the main reasons to implement AI-based solutions?

  • Better customer experience
  • Increased customer acquisition
  • Improved access to financial services for SMEs
  • Fraud detection

 What are the main challenges in developing AI in financial institutions?

  1. 68% of organizations have challenges with the integration with other tools as data is decentralized. Here banks should think about the usage of some of the tools, available on the market (like data customer platform) or do it in-house, to organize the data. When data is organized after that organization is ready to build a machine learning architecture. A good practice here is to build a separate AI module, which consumes the data from the data platform and has communication bridges with other systems.
  2. Integration into company culture and processes (66%). An organization should understand the goals of building an AI strategy.
  3. Lack of Specialized Data Science Team (51%). Such a team can be built in-house (start from a small team) or outsourced (AI as a service is provided by, for example, Asseco Poland in the CEE region).
  4. Lack of trust in AI decisions. Implementation of explainable AI (XAI) can help here. It’s artificial intelligence, which the results of the solution can be understood by humans. It contrasts with the concept of the “black box” in machine learning where even its designers cannot explain why an AI arrived at a specific decision. So, XAI helps to get some insights from the data and get even not obvious correlations, and we can understand the results of data analysis, like why customers received a certain score during the scoring analysis.

Watch the presentation for more details: