AI-driven CRM

Cloud Computing, NLP

AssistantCAT: Contact acquiring and tracking

Addressing the need for more efficient and accessible contact tracking and management, AssistantCAT was started in order to provide voice navigation and more targeted features to help with non-sales related networking. It relies on several AWS applications. This report details the motivation behind the project, the solution architecture, implementation details, and conclusion. At the end is the code documentation and references used.

Problem Statement and Motivation:
One of the biggest drawbacks to mobile phone books or contact relationship management apps is the amount of manual data entry required. To add a person, one has to type name, phone number, notes, history, and so on. While some solutions attempt to automate this with tracking, there is still the manual data entry component in terms of notes to add about a person, which is arguably the most important information in these relationship management apps. AssistantCAT (Contact Acquiring and Tracking) was started to address the need for a more accessible, engaging way to acquire contacts and improve these relationships. One of the defining features is the ability to bypass manual data entry by communicating with a friendly chatbot. With AssistantCAT, instead of manually typing or doing data entry, the user would be able to speak as if they were speaking to a friend. AssistantCAT also provides a couple other features to make networking easier.

Solution Architecture:
As shown in the solution architecture diagram, AssistantCAT uses 9 lambda functions written in Python, two EC2 instances, 1 RDS MySQL 5.6.39 DB instance, an S3 bucket, IAM authentication and Cognito, 1 Lex bot, a javascript API Gateway SDK, SQS, SES, and CloudWatch. Non-Amazon resources include news website, Google, and api.

A portion of this project was the content engine which delivered daily emails to your inbox, with articles recommended for your contacts based on NLP. The backend integrated with, Google Search, and Lex using AWS Lambda.

View Documentation Excerpt (2018)