Chatbot for customer support on METRO Cash & Carry website

Client

METRO Cash & Carry was founded in Germany in 1964. Four years later, the revolutionary concept of a self-service store went beyond the borders of the country. Today, more than 750 shopping centers under this brand operate in 25 countries around the world. The company has about 100,000 employees worldwide.

Context

Today, large retailers are actively introducing digital channels of communication with consumers. Clients and partners want the company to be in touch 24/7 and promptly provide information in all areas — from the availability of goods to the terms of cooperation.

Task

To develop an online assistant on the METRO Cash & Carry website, which will allow:

Provide 24/7 communication between users and the brand and inform customers about the most popular topics.
Automate the call center: optimize the processing of standard requests and reduce the number of requests.
Implement the functionality of the personal account in the bot, giving the user the opportunity to adjust their data by sending requests to the chat.
Announce promotions.
Conduct loyalty programs, brand names, and other promotions with the help of a bot with the loading of checks

Solution

We have developed a virtual assistant with linear logic and a natural language processing module (NLP). Such a hybrid chatbot is as user-friendly as possible: a person can either choose a suitable button or write their own request. After that, he receives a ready response from the database.

The chatbot in the interactive interface gives METRO customers access to all the necessary information. For example, it tells the addresses of stores, informs about promotions and special offers, helps to issue a card, and becomes a partner of the company.

The project took about three months to complete. The main difficulty was the lack of a ready-made dataset with requests from site users. To form it, we generated typical customer requests, and then Metro call center employees helped us to correct phrases and add new ones.

Results

The project is currently in the launch stage.

At the testing stage, the accuracy of recognizing the subject of the appeal was more than 80%.

Come back

Ready to chat?

Do you have any questions? Fill out the form, and our managers will contact you soon.

Cleverbots

Your name

Your phone number

Your email

0/ 1000

By clicking submit,
you consent to the processing of your personal information.