Text analysis of the future: Quality improvement through hybrid AI technology
How TCL.digital was able to get more out of its text analyses using state-of-the-art AI methods and LLMs
As a pioneer in the field of text optimization for marketing and customer communication, Hanau-based TCL.digital combines over 25 years of communication experience with state-of-the-art AI technology. With its Text Performance® method, the company has created an approach that makes language a measurable success factor. With its SaaS platform TEO V3, TCL.digital supports companies in strategically analyzing, optimizing, and using texts in a brand-appropriate manner—for effective communication and reliable competitive advantages.
The project at a glance
- Significant efficiency gains by replacing rule-based systems: Implementation of a hybrid LLM-based approach that replaces the time-consuming development and maintenance of complex, rule-based parser systems, thereby enabling significant efficiency gains.
- Significant uplift in analysis quality: Use of modern machine learning algorithms for semantic text analysis, increasing the accuracy and informative value of the results.
- Greater reliability and risk minimization: Identifying complex and rare text patterns that cannot be mapped in rule-based processes leads to improved analysis and reduces the risk of errors.
- Business value through customer satisfaction and market positioning: More precise analyses improve the quality of text optimization for end customers, increase satisfaction, and strengthen the product's long-term competitiveness.
Background
Parser technologies have been the backbone of semantic text analysis for years. In practice, companies have often relied on highly complex, rule-based systems that are extremely time-consuming and resource-intensive to set up and maintain. Mapping complex sentence structures poses a particular challenge, as these must first be identified and can only be supplemented manually. Despite their high complexity, such systems reach their limits when it comes to reliably capturing linguistic nuances and ensuring precise analysis. Some linguistic subtleties cannot be captured by such systems at all.
The use of modern machine learning algorithms and large language models (LLMs) can significantly improve semantic analysis and enable complex patterns to be recognized automatically. This opens up the possibility of not only efficiently replacing existing rule-based parser models, but also substantially increasing the quality of analysis and creating the basis for future-proof, scalable text optimization processes.
Solution
Hybrid approach
Performance & practicality
Hybrid approach
As part of the project, codecentric worked with TCL.digital to develop a scalable backend system that combines state-of-the-art NLP(1) and machine learning technologies with generative AI (LLMs). This hybrid approach enables significantly more precise semantic analysis and opens up new applications that are not feasible with classic parser technologies—such as the differentiated recognition of different types of imperatives.
(1) Natural Language Processing (NLP) is a subfield of artificial intelligence that deals with the automatic processing and analysis of human languages. It enables computers to understand, interpret, and generate written or spoken language.
Performance & practicality
A key focus was on the performance and practicality of the system: results are delivered within seconds, enabling immediate further processing. At the same time, the architecture was designed to meet the highest standards of quality and robustness while remaining flexibly scalable to seamlessly integrate future enhancements.
The new solution was developed in close consultation with the customer. This ensured that requirements were implemented precisely.
Result
Despite the technological complexity behind the desired functions, a functional software module was implemented within a few weeks. The new system delivered excellent results in various tests. By focusing clearly on pragmatic implementation steps, we worked with the customer to achieve a rapid technical breakthrough that demonstrated the effectiveness of the approach and laid the foundation for the sustainable further development of the system.
Thanks to the new possibilities offered by the hybrid ML/LLM-based system, TCL.digital will be able to offer its customers even more accurate text optimization in the future, thus securing a qualitative differentiator. This will further increase customer satisfaction and secure the company's already leading market position. The new system will be expanded in the future to solve even more complex linguistic challenges. We look forward to further joint development.
Dirk Majchrzak
CEO at TCL.digital
The technological support provided by codecentric has permanently closed a significant gap in our development in the field of machine learning and AI. I was particularly impressed by the solution-oriented approach to NLP problems.
Dirk Majchrzak
CEO at TCL.digital
Any questions about the project?
You are convinced that you can improve your products with AI, but don't know exactly how? Just get in touch with me!
Service Lead GenAI
Further reference projects
Find out about other successful projects that we have completed with our clients. Perhaps you will find ideas here for a use case in your own organization.
Niklas Haas
Service Lead GenAI