Stoen Operator

AI improved information retrieval about easements from paper documents

Stoen Operator Case Study


Our customer, Stoen Operator, used artificial intelligence (AI) to read data from easement agreements and notarial acts. Using AI allowed Stoen to reduce the time needed to manually archive such documents and enter information into databases by over 60%.



Industry: Energy
Country: Poland
Solution: iDoc
Category: Automated document categorization

The company was established in 2007, and currently, it delivers power to over a million recipients in Warsaw and surrounding areas. Stoen Operator’s network includes over 18 thousand underground and overhead cable lines with different voltage levels and nearly 7 thousand energy stations.


Each month, Stoen Operator receives about 150 agreements and notarial acts that must be entered into a database of real estate legal titles. Every document includes about 30 sections that needed to be manually copied into a system. Scanning and copying a single document would take 20 minutes on average.

One person was responsible for copying data from documents, taking them from 1/4 to 1/2 of their full-time job. This work was so repetitive and tedious that people in this position would often quit. Each time, the new employee would need to go through training and learn. Such high employee turnover would increase recruitment spending and lower the process efficiency. Manual data entry also caused a high risk of errors and non-standardized data formats.

Stoen Operator searched for a solution that would improve its process of archiving agreements related to easements and right-of-way. The solution that the company had in mind would simplify and accelerate data entry from legal documents into the internal system, ADX.

Reducing the time required for data entry would allow employees to dedicate their saved working hours to more rewarding tasks. Limiting the time spent on repetitive, tedious work would also lower employee turnover.


Entering agreements into the systems – numbers (before implementing our solution)

the number of agreements received by Stoen Operator each month
number of sections in each agreement that need to be copied
is how many minutes are needed to enter data from a single document


We suggested Stoen Operator automate data entry from legal documents into the ADX system with the iDoc service. iDoc is a SaaS (Software as a Service) IT platform. It uses artificial intelligence and machine learning algorithms to automatically process different documents and their attachments, categorize them, and identify their attributes.

How does iDoc work?

The iDoc system reads and categorizes the contents of scanned pages of legal documents, e.g., notarial acts and different kinds of agreements.

Trained AI models categorize document types and recognize and interpret over 30 kinds of different information. After iDoc processes documents, a user can verify the processed values. If the values are correct, a user accepts and sends the data, including the attached document, into the internal database.


iDoc categorizes documents with at least 96% accuracy and then enters data with about 90% accuracy. The accuracy increases with time as AI models keep on learning. Combining AI with human verification ensures the highest accuracy and effectiveness.

Cooperating with Stoen Operator to implement iDoc for automated data processing, we…

  • completed requirement analysis and verified Stoen Operator’s business processes, analyzing the manual data entry process. We also prepared a list of possible solutions,
  • prepared analytical models and trained them to read the data correctly,
  • we implemented the solution as a web service.

Results and benefits

Implementing iDoc allowed Stoen Operator to…

replace the manual document analysis with the automated process

ensure higher correctness of data placed in the database

accelerate and simplify the document verification

automate manual data entry from documents into the system

As a result…

time required for archiving the legal documents was reduced by about 63%

Stoen Operator eliminated monotonous, manual human work related to copying documents into the system, which lowered the turnover of employees responsible for data entry

the employees can dedicate their saved working hours to tasks related to real estate regulation processes

archivization of documents related to real estate legal titles is significantly faster and several processes, such as the investment process, are more efficient

higher data quality due to the reduced number of errors related to misspelled words

the standardized data format that makes it easier for other systems to use the data. This also eliminated the need for performing periodical database inspections and manually standardizing data

more accurate document categorization provided by AI/ML technologies and significantly reduced time of archiving scans in the database

significantly reduced time of onboarding new employees – iDoc has an intuitive, user-friendly interface

employees develop new skills related to the AI technology

The iDoc system allows Stoen Operator to integrate its ADX database and GIS system. With such integration, employees can access real-estate data on a map.

Entering agreements into the systems – numbers (after implementing our solution)

reduction in time needed for data entry
is how many minutes iDoc takes on average to enter a single document (before it would take employees 20 minutes

Using iDoc allowed us to save 25 out of 40 working hours (63% reduction) spent on archiving documents related to easements and right-of-way. Now, our employees can dedicate this time to other tasks, e.g., real estate regulation processes. Therefore, our investment process is faster, too.

The project was completed diligently, and Globema’s team proved their advanced skills and thoroughness. We consider Globema a reliable and professional contractor as well as a responsible business partner.


Krzysztof Matyjek, Coordinator of the Strategic Real Estate Regulation Team, Stoen Operator

Our cooperation

Stoen Operator started cooperating with Globema at the beginning of the 2000s when the operator was implementing its first GIS system, Smallworld 3.21 version. Since then, we worked with Stoen Operator on many projects in different areas of its operations. Currently, the operator uses several solutions we delivered, e.g., network inventory systems:

  • In 2002 – Smallworld v3
  • In 2015–17 – Smallworld v4
  • In 2023 – Smallworld v5

We constantly work on developing these solutions. We also completed other works for Stoen Operator:

  • Integrating with SCADA, ADX, and CBO systems as well as a task management system, M.TASK
  • Migrating data from the network inventory system Stoen Operator used in the past (about 6 million inventoried devices and objects)
  • Completing a project related to interpreting and entering internal station schemes into the GIS databases

Due to long-lasting cooperation, Globema often partakes in technical and process analyses to improve their effectiveness.

Our cooperation in the area of new technologies

In 2020, we signed an agreement related to a partnership in the R&D project, GlobIQ. Together with Stoen Operator, we worked on improving its different business processes in departments such as Real Estate, Energy Emergency Services, and Asset Management. The processes were related to, e.g., object inspections, investments, outages, and entering real estate legal titles into databases (described in this case study).

During this project, Stoen Operator delivered data and detailed information about the operator’s processes. We attended many meetings with employees from different departments and conducted Consumer Insight interviews to learn about user needs and expectations. We also completed several analyses using data delivered by Stoen Operator. Based on the analysis results, we adjusted tools to the Operator’s needs. In the end, we presented a service that significantly optimizes Stoen Operator’s operations.

One of the GlobIQ project’s goals was to use AI and machine learning to automate data entry from different documents into the company’s databases.

Using machine learning reduces repetitive, manual human work related to data acquisition:

  • categorizing documents
  • acquiring meta-attributes
  • acquiring information from different document types
  • acquiring basic information about the operator’s network

Our cooperation resulted in, among others, the iDoc service for real estate departments (iDoc – Right-of-Way).


Future plans

iDoc is a modern solution that (after adequate adjustments) can be used in different processes. Stoen Operator is planning to implement iDoc in more departments:

iDoc Decisions

supporting the decisions about the right-of-way

iDoc Technical documentation

that fills in the infrastructure device database in the GIS system. Currently, this work is done mostly manually.

Would you like to achieve similar results?

See if iDoc would be a good fit for your company. Fill out the form, and one of our experts will contact you to discuss your needs and schedule a demo.

22 848 73 13

    What do you want to talk about?

    The Controller of your personal data is Globema sp. z o.o. with its registered seat in Warsaw. You have the right to access your personal data, the right to rectify it, the right to restrict its processing, the right to data erasure and data portability, as well as the right to object to the processing of your personal data. Detailed information on the processing of your personal data are provided in the Privacy Policy.

    Read other success stories

    TAURON Group

    Area-based energy forecasting – more effective estimations of prosumer production and consumption

    Learn how Globema’s solution helped TAURON Group forecast production and consumption of energy from renewable sources.

    RES energy production forecasts   | Energy


    Network inventory supports Polenergia’s growth based on M&A

    Learn how we standardized the network management at one of the largest energy distributors in Poland and organized data about 30 areas covering 11 thousand customers.

    EL.GIS / Smallworld | Network Asset Management | Energy

    Schedule a meeting to achieve similar results at your company!