Construction documentation with AI agents

In civil engineering projects such as fibre-optic network rollouts, cec Ingenieure generates large volumes of image and metadata required for verification, documentation and billing. Manually analysing this data is time-consuming and prone to error.

Together with Telekom MMS, cec Ingenieure GmbH has fundamentally modernised this process using an n8n automation solution based on AI agents.

-> Would you like to implement similar AI solutions in your company? We support you from identifying suitable use cases to developing customized AI agents and ensuring productive operation. Discover agent-based AI with n8n.

• Significantly reduced manual effort
• Faster documentation and billing
• High data quality through structured processing
• Scalable solution for growing project volumes

As part of its civil engineering projects in the telecommunications and infrastructure sectors, cec Ingenieure GmbH processes extensive photographic documentation. These images need to be interpreted, classified and assigned to individual construction orders – providing an essential basis for verification, documentation and billing.

Previously, this allocation was largely performed manually. The associated effort tied up personnel resources and significantly increased processing times.

The objective was therefore to automate the entire image analysis process – from capturing the photographs to providing the data in a structured format – with the longer-term goal of fully digitising the process.

The solution needed to process large volumes of unstructured image data, integrate seamlessly into the existing Microsoft environment and enable transparent analysis. At the same time, it needed to be flexible and extensible – from partially automated workflows to more advanced autonomous process steps – in order to accommodate future requirements.

Together with cec Ingenieure GmbH, Telekom MMS developed an end-to-end Agentic AI solution based on the n8n workflow platform. n8n is a flexible low-code platform developed by the renowned Berlin-based AI start-up n8n, enabling complex workflows to be implemented efficiently and in a structured manner. Pipelines can be created via an intuitive graphical user interface and flexibly adapted or extended as requirements evolve.

For the engineering services provider, the image analysis process begins directly on the construction site: photographs are taken on site and subsequently uploaded to a central project platform. From there, they are automatically processed in a structured workflow, with the images sorted overnight, enriched with data and made available for further use.

A key component of the solution is structured photographic documentation using metadata. Instead of conventional folder structures, flexible metadata fields are used, enabling targeted filtering and sorting as well as rapid access to relevant images.

During AI-assisted processing, each image is automatically analysed and enriched with project-relevant information. This includes, among other things:

• Location data such as the POP area, network distribution area or address
• Content classification according to construction phase, component or route type
• Additional information from EXIF data, such as the time the photograph was taken or the responsible person

One specific use case is automatic localisation through GPS mapping: based on their coordinates, images are assigned to the nearest building or construction section. This completely eliminates the need for the previously manual allocation process.

Another example is AI-based image classification: a model specifically trained using Azure Custom Vision identifies typical objects and work steps in civil engineering – such as main routes, building connections or specific components – and automatically assigns each photograph to the most likely appropriate category.

With metadata-based documentation, each image is enriched with structured information, ensuring that it can be clearly identified, retrieved and analysed at any time.

The workflow also automates additional processing steps, including extracting image information such as EXIF data, enriching images with project-related information and optimising file sizes for efficient storage.

The solution is seamlessly integrated into the existing Microsoft system landscape – particularly SharePoint for structured data storage and Power BI for analysis and reporting. This creates a consistent, scalable data foundation for documentation, analysis and billing.

The n8n platform connects systems, data and AI models within clearly defined workflows, creating the foundation for controllable agentic processes. This makes workflow automation transparent and verifiable.

Thanks to the individually developed workflow and standardised process, image data at cec Ingenieure GmbH is now automatically captured, classified and assigned to the relevant construction projects. Structured processing ensures consistent and reliable data – an essential foundation for documentation and billing.

cec Ingenieure GmbH also benefits from significantly faster processes and a noticeable reduction in employees' workloads by reducing manual effort. Resource-intensive routine tasks are automated, leaving more time for value-adding activities. The AI agents therefore contribute to lower costs, provide competitive advantages and support higher-quality decision-making. Data protection and governance requirements are fully observed.

In addition, automated image processing creates transparency across all project data, reduces potential sources of error and enables reliable scaling as project volumes increase. At the same time, the agent-based solution lays the foundation for further AI-supported automation steps.

cec Ingenieure GmbH provides engineering services for the telecommunications and energy supply sectors. The company supports municipalities and other clients from planning through to documentation with integrated digital services – ranging from approval and execution planning to the implementation of complex fibre-optic and power infrastructure projects.