Every year again: The meter reading must be submitted to the energy supplier. This is an annual routine task that can be made easier for both customers and customer service agents of energy companies. Because often a complicated process lies behind this task. Some energy providers advertise with customer portals that allow data entries after a user login. Others only offer the telephone channel. In most cases, however, this transmission process is frustrating and time-consuming for the customer: Either passwords are missing or they are stuck in a waiting loop on the phone for a long time.

It does not have to be so complicated and elaborate:

For many customers, direct contact with employees is no longer the main criteria for excellent customer service. Ultimately, they simply want to have their concerns resolved as quickly as possible. And they prefer to use self-service solutions.

"76% of consumers globally prefer to first try to solve issues on their own before contacting support."(freshworks, 2019)

Chatbots are ideal for the implementation of self-service solutions. This technology can be integrated into live chats or channels like Facebook Messenger or WhatsApp. There they automatically answer customer queries 24/7 with the help of artificial intelligence. They are also suitable for the use of process-oriented conversations.

One example of how a chatbot can be used as a self-service solution is the chatbot Karl from the energy supplier Süwag. With its help, customers can quickly and easily submit the meter reading on the website in a few automated steps.

How exactly this works and what to look out for when developing such a chatbot use case is described below. Please note that the chatbot is only available in german. Screenshots are therefore not available in english. 

Step 1: Welcoming the users

Chatbot Karl

The chatbot should best inform in the first messages about the tasks it can perform and the questions the user can ask. Preset buttons can be very helpful in this respect, as they guide the user through the process and speed up the input.

Step 2: Date of meter reading

At the beginning the users are asked for the date of the reading. Here it is important to check the logic in the background via code to ensure that the meter reading period is correct and, for example, is not in the future. Often the meter reading period must not be older than 14 days. In addition, the chatbot should be able to understand and process user input such as "yesterday" or "today a week ago".

Step 3: Identifying the customer

In order to be able to assign the transmitted meter reading to the correct customer, these must be identified in advance. Which data must be requested for this depends on the company. Usually information such as customer and meter number is required. Users should be informed as clearly as possible what information is required and where they can find it. For example, the chatbot can indicate how many digits the customer number contains and that this number can be found on all letters from the company to the customer.

In addition, the chatbot should be able to continue the process if the customer does not have the required information at hand. For example, the chatbot can be asked for alternative means of identification such as address, birthday, etc.

Step 4: Enter the counter reading

The chatbot was previously connected to the company systems by means of integrations. This means that the information entered can now be used to check in the background which tariffs the customers have chosen (e.g. gas, individual tariff, high tariff). Based on this information, the corresponding meter status can be requested. Here again, it is important to communicate as clearly as possible how the meter reading should be entered: With or without decimal place, point or comma, etc. As a helpful orientation, the last meter reading can be displayed, for example.

An even more comprehensive user experience can be offered to customers through the integration of OCR technology. Through this the bot is able to extract information like the counter reading from an image. Customers only need to upload a photo of the electricity meter and the manual entry of the numbers is no longer necessary.

Step 5: Confirmation

After the counter reading has been transmitted and successfully saved by the system in the background, the user can be offered to receive a confirmation e-mail. This written confirmation not only confirms that the current meter reading has been successfully transmitted, but also strengthens the users' confidence in the Self - Service Solution.

Step 6: Handling transmission errors

In the event that the entered value cannot be transmitted to the system, the work of the users:inside should not have been in vain. To ensure customer satisfaction, we therefore recommend that the information collected be forwarded to a customer service representative for problem solving.

Step 7: Dealing with unusual values when entering the meter reading

Now and then the indicated counter value deviates unusually from the last value. The reason for this can be a change in consumer behaviour or simply a typing error during entry.

Therefore, in such cases, an explicit question should be asked about the reason for the discrepancy. It also makes sense to give the user the possibility to re-type their counter reading and correct it if necessary.

Step 8: Feedback

Last but not least we think it is essential to ask the users for feedback. This is the only way to measure whether the chatbot leads to happier customers. For the collection and analysis of these data the conversational platform BOTfriends X. Using this platform, conversation content can be continuously adapted without programming knowledge and important KPIs (Key Performance Indicators) can be measured.

What else?

If you want to learn more about chatbots in the energy industry, you can download our case study from Süwag. Learn more about the project approach and the general Conversational AI strategy.