KIK — Service Analytics
Intelligent Analysis of Customer Service Data
Together with Humboldt University of Berlin, we are developing new AI methods to not only answer support requests — but to deeply analyze them.
Customer service data is one of the most valuable — and least utilized — sources of information in businesses. The research project KIK makes this data systematically usable.
In the KIK project, we are developing new AI methods together with Humboldt University of Berlin to extract valuable insights from daily customer interactions: Which products are causing problems? Which topics are trending? How do service contacts impact business outcomes?
The results are made available via an open API for web tracking systems such as Google Analytics — creating a bridge between customer service and marketing that has never existed before.
The Problem: Service Data Remains Siloed
While marketing and web tracking data are analyzed in great detail, service contacts often remain a blind spot in the customer journey. Companies face questions they cannot answer today:
KIK closes this gap: By systematically integrating and analyzing service data in conjunction with web tracking systems, a comprehensive picture of customer behavior emerges — from the first website interaction through the service contact to the purchase decision.
Our Four Research Goals
Entity Recognition in Service Data
Automatic recognition and linking of concrete entities — products, features, issues — in customer service interactions. For a far more granular analysis than conventional topic-based evaluations.
Topic Grouping & Trend Detection
Automatic categorization of service requests and early detection of emerging trends. Enabling companies to identify problems before they escalate — and act proactively.
Anomaly Detection & Forecasting
Predictive models based on modern deep learning methods establish baselines for service contact volumes. Significant deviations are automatically detected — an early warning system for customer service.
Integration with Tracking Systems
An open API connects service analytics with web tracking systems such as Google Analytics. For the first time, service contacts can be linked with user behavior — enabling true root-cause analysis.
The Partners
OMQ GmbH
OMQ is a Berlin-based AI company that has been developing intelligent automation solutions for customer service since 2012. With over 250 international customers — including Deutsche Bahn, Tchibo, and Mister Spex — OMQ processes thousands of service requests daily through chatbots, email bots, and other channels.
In the project, OMQ takes on the central role as integrator: service data is collected, analyzed using AI-based methods, and fed into existing tracking systems.
omq.ai →Humboldt University of Berlin
Two renowned research groups from HU Berlin participate in the project, contributing their scientific expertise in key areas of artificial intelligence:
Machine Learning Lab — Prof. Alan Akbik
State-of-the-art methods in natural language processing and entity recognition. The team is responsible for developing robust models for recognizing and linking entities in customer service interactions — even in specialized domains with very limited training data.
Information Systems Lab — Prof. Stefan Lessmann
Extensive expertise in predictive modeling and time series analysis. The team develops deep learning-based methods for forecasting and anomaly detection in service data, with rigorous scientific evaluation.
Key Facts
Co-funded by the European Union
This project is co-funded by the European Regional Development Fund (ERDF) and supported under the IBB ProFIT program (Program for the Promotion of Research, Innovation and Technology) by the Investitionsbank Berlin (IBB).
The State of Berlin and the European Union support applied research and development of innovative products, processes and services in Berlin through IBB ProFIT — strengthening the innovation ecosystem and competitiveness of the regional economy.
More about IBB ProFIT →