What are the levels of artificial intelligence

Artificial intelligence: This three-level model is intended to systematize AI in the financial sector

Twelve companies in the financial sector are currently developing a structured approach for the use of artificial intelligence over a period of six months. Now the consortium project “Artificial Intelligence in Financial Services” with the participation of the software company PPI has presented its first results. With the help of a new three-level model, financial companies are given the opportunity to systematically advance to the appropriate AI process.

There is hardly a topic that is currently generating so many hopes for increased efficiency and a paradigm shift in the economy across all industries as Artificial Intelligence (AI). Many companies attribute immense future potential to AI - also and especially in the environment of financial companies and insurance companies. In order to use AI in a really targeted and value-adding way, a systematic approach is essential. Together with the ibi research institute and other partners, PPI AG has developed a three-level model for determining needs, implementing and using artificial intelligence at financial service providers. The model was created as part of Germany's most comprehensive consortium project on AI in the financial sector. In addition to PPI AG, the companies involved included Datev, Deutsche Leasing, Targobank and Fiducia & GAD IT.

Artificial intelligence: systematically using potential

The three-level model brings a scientifically developed approach into practice. For its validation, use cases from all banking-related areas - such as consulting, service or payment transactions - were used in the course of the project. The model enables financial companies that want to use artificial intelligence to systematically identify the appropriate AI process from the business process via defined application classes.

The upper level describes the Business process: What does the business process to be addressed look like and how can its consideration be focused on the relevant steps? Today's areas of application for AI are often in the areas of marketing, sales, and risk and fraud detection. In the future, however, widespread use in financial companies is to be expected.

The middle level describes Application classes or standardized Task typesthat are hidden behind the relevant steps of the business process with its functions and data and that can be supported by AI. One possible application class is the examination of the task context in which the artificial intelligence translates a customer request for old-age provision entered informally on the bank's website into the bank's portfolio and then offers the customer, for example, ETF-based asset management.

The lower level helps with Allocation of the appropriate AI procedures and their respective characteristics for the application classes. For example, if the input transformation application class involves translating speech or text into digital code, deep learning using neural networks is an option.

The collaboration of experts from different parts of the financial value chain was extremely productive. With the three-level model, we have developed an efficient instrument to take the topic of artificial intelligence to a new level together with our customers. "

Stephan Heinze, Consultant, PPI

Goal: Holistic understanding of AI processes

AI is already part of many customer projects at PPI. The consultants and software developers also consider topics such as chatbots, fraud detection or robotic process automation. “In the future, we want to give our customers even more support in understanding AI holistically, planning and using it. The consortium project gives us significant impetus here. "

A detailed description of the three-level model is contained in the whitepaper "Artificial intelligence in financial services - from the analysis of application scenarios to the development of a framework for the use of AI". You can download it free of charge. partly

You can find this article on the Internet on the website:

(5 Votes, average: 5,00 of a maximum of 5)
Loading ...

Interesting too