Automated processing of energy bills
The following scenario is familiar to most: Once a year, an employee of an energy service provider walks through the basements of the properties in the neighborhoods and reads the meters and measuring devices for electricity, gas and water. This then results in the energy bill, which lands in the tenant’s mailbox. The situation is similar, but in a much more comprehensive version, for energy bills from real estate management companies, for example, or from SMEs as consumers. Extended selection options and significantly higher bill amounts distinguish private from business customers. At the end of the day, however, they all want the same thing – to see the best possible amount on their energy bills. In this article, we’ll take as an example a property management company and how it handles energy bills, and how an intelligent, AI-based OCR (Optical Character Recognition) can provide valuable insights for comparisons.
Energy bills primarily have the same data as traditional invoices, including supplier data, invoice recipient, invoice and customer number, invoice amount, and so on and so forth. What is special about energy bills, however, is that they do not show conventional item data, but rather work with meters, time periods and the number of services purchased. This means that an energy bill always shows the period to which the data refers. The meter, which is registered under a number (i.e. own for electricity, gas and water) is followed by the meter reading X old and meter reading X new, which refer to the noted billing period. Finally, a factor is used to calculate the number of purchased services in kilowatt-hours (for electricity and gas) or in cubic meters (for water), which in turn results in the individual settlements. With an emphasis on individual settlements: Depending on how much service is purchased from an energy service provider, the electricity product can in fact be procured on the free market. From an annual consumption of at least 100,000 kWh of electricity (with some energy service providers also from 50,000 kWh), the origin and composition of the electricity can be more or less freely selected and purchased, i.e., for example, a green electricity product that is composed of sustainably produced electricity from hydropower or even additionally from solar energy. As a low-cost alternative to renewable energy, electricity can also be selected from fossil fuels such as natural gas. Due to the different compositions, the selection spectrum is therefore large. What is already tedious to read, is also tedious for the people who finally have to process and compare these energy bills. Real estate management firms with their sometimes numerous properties, scattered throughout the country or even in various countries, logically often obtain the energy for their properties from different suppliers, which means that they have to deal with a large number of different energy bills for processing. In other words, they have to deal with highly unstructured data in different formats. However, since the range of offers on the free market is large and comparing the energy bills of individual properties with those of individual providers provides valuable insights into consumption, they depend on specific data. This is the only way they can put the energy service providers through their paces and compare where the electricity or gas costs how much so that they can optimize their contracts if necessary or consider switching.
The solution here for simpler and clearer handling of all energy bills is called intelligent, AI-based OCR (Optical Character Recognition). The OCR software automatically extracts all relevant data, i.e. time periods, meter readings, services, consumption and above all – very important for comparing energy bills and energy service providers – the compositions, proportions and price variances of the electricity or gas purchased. The OCR automatically captures and extracts the unstructured and discrepant data from the energy bills and prepares it in such a way that the data is now structured and can be used by employees in a simple way for specific comparison. In this way, the contract can be optimized or changed if necessary. What applies to the application of intelligent, AI-based OCR to energy bills of a real estate management company as a consumer can, of course, also be used in a similar context between electricity producers, whether hydroelectric or nuclear power plants and electricity suppliers.