Building Conversational Bots Using Amazon Lex

Sachar Gaming wanted to partner with Searce in order to leverage its long experience with solutioning on Amazon Web Services infrastructure and build a reliable chatbot. Searce has been building chatbots utilizing AWS services & helping partners from all over the world.



  • Sachar Gaming is one of India's largest online gaming & E-commerce firm
  • The firm specialize in multiplayer, real-time gaming
Sachar Games objective to create a ChatBot to Share the Additional Workloads
  • Increasing Service Requests - With the increase in the number of game offerings & user base Sachar Gaming is experiencing heavy loads on the support teams. This is mainly due to a high number of service tickets and a limited number of team members to serve the request.
  • Scalability of workloads - As the user base increases the number of parallel users rises resulting in high demands of infrastructure. The ChatBot needs to be built on a platform which supports scaling as and when necessary.
  • Control Costs - Sachar Gaming wants to keep a check on costs by limiting the support team size. To be able to achieve this there is a need for a ChatBot which can work in parallel to the existing support flow & share the additional workloads.
  • Improved Customer Experience - Due to longer waiting times, end customers may face bad experiences. Through this implementation, Sachar Gaming wants to overcome any such possibilities & provide customer support with a user-friendly experience.
  • Collaboration with the current Process - Sachar Gaming already has a mature support process setup, which uses third-party tools. The challenge will be to be able to build a chatbot which can be functional without disturbing the current flow of processes.
Building a Solution using Angular JS

Team of architects from Searce had a detailed discussion to understand the entire flow which exists in the organization for support activities. The discussion was converted to a diagrammatic representation for the scope of this project.

Searce created multiple variations in utterances on any statement made by a customer and trained them on Amazon Lex. Lex was not trained to respond according to the input command in simple English language. This training was performed in iterations along with tests to improve the results.

The chat window was built using Angular JS & the connecting backend layer between Front End & Lex was built using Python Flask.

The final chatbot could handle client request like -

  • Payment issues
  • Login issues
  • Fund transfer
  • Status claim
The Right Chatbot to Improve Customer Experience with Reduced Manual Efforts
  • Reduction on Calls - Implementation of ChatBot has increased the customer time on the mobile platform when compared to voice time using contact centre options
  • Increased Customer Satisfaction - Having fewer calls, reduces the customer waiting time which results in an improved customer satisfaction
  • Ticket Creation - Manual effort of creation of tickets on Zendesk, after listening to the calls was now done using ChatBot