Client:
Water companies
Our role:
We developed and implemented a fully automated AI solution capable of accurately analysing urban non-absorbent surfaces. This identified potential discrepancies in client-reported data.
Country:
Slovakia
Setting the scene
Water companies face challenges in accurately calculating rainwater treatment fees. This is due to a reliance on client-reported data about impervious surfaces, such as roofs, roads and car parks. Precipitation falling on these surfaces ends up in the sewerage system and requires cleaning along with wastewater.
Accurate information about these surfaces is important because it dictates the fees ultimately charged for rainwater treatment. Due to the extensive coverage of water company networks across urban areas, physical checks of each plot are impractical. This means water companies typically rely on potentially inaccurate, client-reported data that often underestimates the actual treatment volume.
How we helped
Our approach initially involved using aerial photographs and satellite images for remote sensing. While visually inspecting large areas manually isn't feasible, AI allows for automated, rapid analysis of hundreds of square kilometres.
We developed a solution using a computer vision model to differentiate surface types from aerial photographs. Outputs from this model were then cross-referenced with land registers and water company data, identifying discrepancies between the reported data and actual conditions. This enabled water companies to focus on inspecting areas with suspected inaccuracies, rather than assessing data manually.
Additional automation steps included data quality control and automated communication with our clients regarding discrepancies, complete with visualisations and inspection results.
"Our AI solution revolutionised how water companies verify impervious surfaces, resulting in substantial cost savings and efficiency gains. By automating the inspection process, we helped our clients to focus their resources on genuine discrepancies rather than manual inspections."
Karsten Hegel, Partner, CEE Export Leader, PwC SlovakiaImpact and potential
This project showcases the transformative potential of technology in resource management. Our analysis found that up to 60% of data on impervious surfaces was underestimated, leading to rainwater treatment fee shortfalls of up to 40%.
In a regulated pricing sector, this forced our clients to cover these unmet costs. Our tailored solution not only addresses rainwater treatment payments but also allows for broader inspections of unreported surfaces and public areas. The integration of AI technology promises long-term improvements in accuracy and operational efficiency for water companies.
Contact us
