tetrel Logo tetrel Logo

Process digitization with innovative text recognition

Novec uses tetrel’s docuduct Platform to extract important information from free text and innovate processes.


Dedicated teams at Novec support cell tower owners, customers and municipalities. In a high frequency communication environment, relevant information must be readily available for team members to ensure efficient communication and high customer satisfaction. New knowledge must be documented quickly and with high precision for operational efficiency.

Free-text conversation notes, typed by an agent during a call, contain important information about cell tower sites, contact persons and the status of building measures and acquisitions. With a growing amount of information, finding the relevant details quickly becomes harder and harder if they are not continuously extracted and saved in a structured format.

Structuring information with tetrel’s Text Intelligence Platform

Novec and tetrel collaborated to launch a new workflow tool, integrating structured information from three years of historical data into an easily usable interface for Novec’s agents.

To extract structured data from conversation notes, tetrel trained a deep learning based docuduct Module to recognize custom information objects relevant to Novec in free-form texts. By using custom training technologies, an accuracy of more than 98% was reached just two weeks after the start of the project.

With a new custom workflow tool specifically optimized for Novec’s use case, agents can easily access all relevant information in their daily work and document new knowledge quickly and effortlessly, leading to faster processing times and a significant reduction in error rates.


reduction in average processing time for tasks


reduction in processing errors


KPI dashboard available to management