The use of AI in construction financing for private customers
Artificial Intelligence (AI) has already made inroads into various domains. The integration of machine learning into construction financing can bring advancements in areas such as advisory, property assessment, and financial security for lending. The current trend in information technology is dominated by AI and its subfields of machine learning and dynamic automation. This paper proposes a scenario for the transformation and automation of the construction financing industry through AI and provides condensed recommendations for action.
The outcome highlights the need to embrace the trend and reap the economic benefits of AI’s utilization
Introduction
This study focuses on the German Bank Institute Sparkasse to examine the impact of AI in construction financing. The objective is to address the question ”What is the influence of AI in construction financing?” and determine the direction that Sparkasse should take if it intends to utilize AI and machine learning technology.
The existing process for construction financing involves consultation in a branch and multiple appointments, resulting in a statement on the feasibility and cost of financing, and the conclusion of a loan agreement if approved. Sparkasse already offers online banking services for basic transactions and account management. The research examines the potential for expanding online banking to offer remote advisory services to customers. The study is based on the author’s Master Thesis.
Conceptual foundations
Bank Institute Sparkasse
The Sparkasse is a bank widely used in Germany, organized under public law in its own financial network. This association describes itself as follows: ”As financial institutions under public law, they — together with their alliance partners from the Sparkassen-Finanzgruppe — fulfill a uniquely responsible mission: they promote the prosperity of the people and the growth of the economy with their products, services and local presence. Locally, regionally and nationwide.”[1]
Use of IT at Sparkasse
The financial informatics (FI) division, operating independently, manages and advances the IT infrastructure of the Sparkasses within the umbrella organization. The in-house development department has designed, developed, and implemented the OSPlus neo software, used internally. The FI is also responsible for the development and maintenance of the individual Sparkasses’ websites, including hosting and providing interfaces for online banking for customers. The utilization of mainframes is a long-standing tradition at the FI and will be maintained through the transition to cloud computing.[2].
Basics of AI and Machine Learning
Artificial Intelligence (AI) is a subfield in computer science that focuses on the automation of intelligent behavior and machine learning. AI is further categorized into two types: strong AI and weak AI. Strong AI is a theoretical concept that aims to perform tasks at par with human capabilities. On the other hand, weak AI is designed to address specific application problems. Additionally, AI encompasses several sub-domains, such as logical reasoning and approximation methods.[3]
The Fraunhofer Institute defines machine learning as follows: “Machine learning (ML) aims to generate ‘knowledge’ from ‘experience’ by using learning algorithms to develop a complex model from examples. The model, and thus the automatically acquired knowledge representation, can subsequently be applied to new,potentially unknown data of the same kind.”[4]
Usage of Chatbots and Voice Recognition
Chatbots are a type of bot. Bots can be defined as follows: “A bot is a computer program that processes recurring tasks largely automatically. Examples that could benefit from machine learning include chatbots, social bots, and gamebots.”[4]
Prominent chatbots widely adopted include Apple Siri, Google Assistant, and Amazon Alexa. These chatbots are equipped with voice control and recognition capabilities and are capable of generating commands through voice input.
The Fraunhofer Institute defines Natural Language Processing (NLP) as the technology for interpreting and generating natural language in spoken and written form. This encompasses speech-to-text conversion, sentiment analysis, information extraction from text, machine translation, and conversation. NLP is utilized to process and analyze language for computer systems.[4]
Case scenarios
Factual scenario
In construction financing, the borrower and the lender engage in a financial transaction involving the financing of a property. The borrower may be an individual or a family, while the lender is typically a bank, such as Sparkasse. To assess the creditworthiness of the borrower, a private company called “Schutzgemeinschaft für allgemeine Kreditsicherung” (Schufa) is utilized. The financing process involves the participation of human actors, computer systems, and various processes.[5] In the study by Price Waterhouse Coopers, titled ”Der große BauFi boom”, conducted in February 2020, the share of construction financing in Germany was analyzed among other debt financing products offered by Sparkasse. The results showed that in 2019, the volume of construction financing in Germany reached 1.3 trillion, constituting 42 % of all customer loans.[6]
Construction financing is a product offered by the Sparkasse and is one of the various debt financing options. The transformation of the state from ”No construction financing” to ”One construction financing” involves a series of processes, which can be considered as one comprehensive process and can be broken down into smaller processes. These processes include property and object evaluation, consultation, and contract discussion.
During the consultation, the feasibility of construction financing and the necessary requirements are determined. The appointment is arranged to emphasize the binding nature of the potential construction financing. Necessary documents are discussed, and the actual contract discussion takes place separately. During the consultation, the borrower’s income is disclosed by querying Schufa and submitting salary statements. This enables the Sparkasse to assess the feasibility and amount of the construction financing requested.
The property and object evaluation is carried out to ensure the loan is not overpriced and that financing is not a losing proposition for the lender. A positive outcome of the processes is the successful conclusion of the contract.
Target scenario
The use of weak Artificial Intelligence (AI) in construction financing enables a borrower to conduct the entire process, from initial consultation to contract conclusion, online and in a single session. The AI utilizes the previously collected data and assesses creditworthiness, property to be financed, among other factors, using sources such as Schufa. This automation reduces the dependence on manual work and the need for physical appointments with bank employees. The use of voice control and recognition technology also allows the borrower to interact with a chatbot during the consultation. This increased flexibility and efficiency provide the borrower the opportunity to complete construction financing at their convenience, without being restricted by the bank’s operating hours.
Interviews
In a study, six IT professionals were interviewed, consisting of two project managers, two software developers, and two IT support staff. Three of them were employed in the banking sector, while the others were employed in non-banking environments. All of them had some level of prior knowledge and experience with AI. The results showed a dichotomy in their perception of AI in the banking sector. Those working in the banking sector primarily viewed AI as a means to enhance security, while those in non-banking environments emphasized the importance of involving employees. However, both groups agreed that the integration of AI in online banking, allowing for decoupling of construction financing from branch hours, was a positive development.
Summary
Operability of AI
The utilization of AI in the banking sector, specifically in construction financing offered by Sparkasse, holds significant potential for growth and expansion. The implementation of AI can allow for further digitization of the online branch, providing customers with increased independence and accessibility to Sparkasse’s services. AI-based evaluation of customer-submitted documents and real estate structures can bring efficiencies in the processing of construction financing applications. The use of AI-powered chatbots as virtual advisors further reinforces the integration of AI in the banking sector.
Data Protection
The Sparkasse faces significant legal and business requirements for data protection and security, including compliance with the General Data Protection Regulation (GDPR). The secure handling of personal data is critical to the company’s operations and business model. To ensure data protection and security, robust measures have been implemented in the online branch, and these measures must be extended to the chatbot or virtual advisor. When handling customer data, proper anonymization or omission of sensitive information must be employed, as per regulatory requirements. All customer-provided data processed by the chatbot must be properly classified and treated as confidential and personal.
Conclusion
The implementation of AI technology in Sparkasse’s processes has the potential to have a significant qualitative impact. To realize this potential, the AI system must replace existing evaluation systems.
Sparkasse’s IT department, Finanzinformatik, is responsible for centralizing the bank’s IT operations, including the introduction and implementation of AI technology. The primary challenge in implementing the AI system is not only to build the system itself, but also to develop seamless integration with the existing systems.
Given these factors, it is important to consider that full automation and positive economic impact from the implementation of AI technology will likely take time to materialize. Implementation of the AI system may require 6–9 months, and staff reductions may only be realized after the introduction of the system to private customers. Consequently, the impact of AI technology on Sparkasse should be viewed in the medium term.
References
[1] Sparkassen-Finanzportal GmbH (2020): “Die Sparkassen-Finanzgruppe: Wir ”uber uns” https://www.sparkasse.de/service/wir-ueber-uns.html, Sparkassen-Finanzportal GmbH
[2] Detlev Klage (2015): “Cloud Computing in der Sparkassen-Finanzgruppe”, S. 1, https://www.f-i.de/content/download/21235/274574/file/ bf6436dc2c93383fbe048787c8946128.pdf, gi Geldinstitute
[3] T. Ramge (2018): “Mensch und Maschine”, S. 19, Reclam
[4] I. Döbel et al. (2018): “Maschinelles Lernen — Kompetenzen, Anwendungen und Forschungsbedarf”, Bundesministerium für Bildung und Forschung
[5] SCHUFA Holding AG, https://www.schufa.de/schufa/unternehmen/so-funktioniert- schufa/
[6] T. Rederer (2020): “Der große Baufi-Boom”, , PwC GmbH WPG
