Business Requirement
The client required a chatbot solution that would serve as a sales assistant that helps shoppers by guiding them through the search process. If shoppers are not able to find what they're looking for using the chatbot, the discussion is forwarded to a live agent.
- Promptly attend to customer queries and resolve them without delay.
- Facilitate decision-making and order placement for customers.
- Improve overall accessibility of product information.
- Update customers about their orders and help them track its movement.
- Enable third-party integrations and facilitate in-depth reporting.
QBurst Solution
We leveraged Dialogflow (NLP platform) and Spring Boot (Java-based framework) to create the core solution.
The conversational flow was designed using Dialogflow to guide users through specific processes, such as making a purchase. Dialogflow uses machine learning algorithms to understand and interpret user inputs. Numerous training phrases were used to handle product search and purchase scenarios. Additionally, conversational responses were reviewed and evaluated to ensure accuracy and consistency. The process also involved manual testing with a variety of inputs. Dialogflow's training and testing features were leveraged to evaluate chatbot performance and identify areas for improvement automatically.
We used the Dialogflow API to integrate the solution with Spring Boot. This enabled us to use the functionality of Dialogflow within Spring Boot microservices. Spring Boot microservices communicate with the Dialogflow agent by sending and receiving data through the Dialogflow API. Dialogflow API handles NLP by interpreting user intent and uses the Spring Boot microservices to provide the actual responses.
Microservices use the information provided by Dialogflow to generate responses and relay the responses back to the user through the chatbot interface. In addition, microservices are also used to communicate with live agents, enabling the chatbot to escalate to a human agent when necessary.
BigQuery is used to store and manage the data generated by the chatbot, such as user interactions, responses, and other relevant information. This data is used to analyze the performance of the chatbot, identify areas for improvement, and provide insights that can help make data-driven decisions. In addition, we leveraged Dialogflow tools and features to manage and improve chatbot analytics, logging, and integrations with other systems and services. This enabled us to create a seamless user experience and optimize the chatbot solution over time.
Additional Cost Savings with Smart Implementation
Originally the client had to pay the managed service provider (MSP) for every conversation even if an agent was not required. Therefore we added an interfacing middle component in the implementation which helped the client to save on cost by paying the MSP only in case of agent involvement.