KIMaKu

The current uncertainties on the global markets and the associated dynamics pose enormous challenges for retail customers today. Price fluctuations or difficulties in delivery result in changes to customer orders at short notice. In order to manage the volume of customer inquiries associated with these dynamics, the aim of the KIMaKu project (AI-based management of customer inquiries) is to automate customer service by means of AI-supported analysis of these inquiries.

To this end, customer needs are to be extracted from the incoming inquiries, and corresponding recommendations for action have to be derived in a fully automated way. As a result, both, customer satisfaction and the attractiveness of the trading platform can be increased simultaneously.

As part of the project, fortiss is researching the use of machine learning methods to process text data. In addition to recognizing entities and assigning queries to categories, the main focus is on the representation forms of the text components for further processing.

Both classic clustering and classification methods as well as language models (large language models, LLMs) are used. The methods will be augmented by the integration of external knowledge sources such as expert knowledge, ontologies and knowledge graphs (hybrid AI).

Funding authority: Bavarian Ministry of Economic Affairs, Regional Development and Energy (STMWi)

BayVFP funding line Digitalization, Information and Communication Technology, IuK-Bayern,
DIK-2308-0036//DIK0547/01

Project management organization: VDI/VDE Innovation + Technik GmbH

01.01.2024 - 31.12.2025