Customer Service Chatbot
For Shared Mobility Solutions





About Client

A European technology company providing software solutions, services and
enterprise-grade shared mobility solutions to the automotive industry.

The client contacted Azilen Technologies to develop an automated response system that can handle inquiries generated via its website, mobile app, phone calls and email. The client envisioned having a single-point solution that can efficiently manage inquiries and automate the responses for better connect.


The expert team at Azilen brainstormed ideas to develop a single-window automatic response solution that can handle a large volume of interactions. The team realized that three elements will have to be handled and the following challenges will have to be tackled via technology.

  • Simplifying a complex business process
  • Improving the customer interaction process
  • Streamlining interactions across multiple touch points


The complex requirement and current As-IS condition prompted Azilen to devise a unique solution in the form of chatbot integration to automate the process flow. Comparative analysis of leading technologies was performed, and RASA stack was chosen owing to multiple benefits offered by the platform.

After defining the use case in detail, performing statistical analysis, integrating slack as channel and creation of development environment, the bot development process was segregated into multiple stages including data preparation, data processing, intent detection and entity detection stage.

On successful development, the chatbot was integrated into the existing process and started functioning by responding to queries. The bot was able to easily reply to questions and connect the prospect to the relevant person or department, when required. Overall, the chatbot was enriched with the following learning capabilities:

  • Direct Supervised Learning The bot was made adept through defined stories that made it equipped with capability to handle basic queries right from the beginning.
  • Indirect Supervised Learning Machine Learning algorithm was used to improve the bot on-the-go via generalization to new dialogues.

  • Self-learning capabilities The bot was embedded with self-learning capabilities to build intelligence via continuous interactive learning training. This included:
    • Developing an understanding of interactions with end-customers and learning from the interactions.
    • Monitoring of responses and feeding back the actions to the knowledge base
    • Flagging and feeding of any errors that emerge into the Known Error Database (KED)
    • Initialization of interactive learning that will make it possible for the chatbot to logically handle queries where stories, intent and actions are undefined.

Over time, the chatbot became capable of learning, generating, predicting and recommending automated responses to the customers. The success of chatbot started the process of email bot and voice bot development that were also successfully integrated in the next stages.

Tools & Technologies

  • - Python
  • - RASA Stack
  • - Tensorflow
  • - Textblob
  • - Python
  • - RASA Stack
  • - Tensorflow
  • - Textblob
Benefits of Automated Response Capability Integration
  • Improved customer service efficiency
  • Reduction in redundant inquiries leading to higher focus on critical inquiries
  • Central channelization of process and efforts
  • Enhanced customer service experience
  • Personalized customer connect due to immediate attention and reduced wait time