Skip to content

sameer-goel/Smart-HR-Bot-with-SentimentAnalysis-Integrated-with-SMS-and-other-services

Repository files navigation

Amazon Lex HR Portal- Smart Bot with Sentiment Analysis integrated with SMS and other services

Build Amazon Lex chatbot to interact with Lambda, Amazon DynamoDB, Amazon Pinpoint, Amazon Comprehend, Amazon Connect, Twilio, Slack, Facebook, Kik and Alexa devices.

Architecture

architecture

Motivation to build this bot

Bot is built for those employees (say warehouse field workers) who does not have internet access and line up to a paper based register, to log basic daily tasks. With this bot they can do it using SMS, Call, Ask Alexa as regular dialog coversation like you are asking your assitant or friend to do the task for you.

The task this bot can do are:

  1. Logging daily work hours.
  2. Calculate total pay for the month.
  3. FAQ search like polcies, agent, salary etc
  4. Most importantly, all the interactions are beging analyzed by Amazon Comprehend for sentiment analysis. This way if the user is not happy we can divert the interation immidiatly to human. You can always add more more HR functionalities like vacations calander, leave balance etc

field_wrorkers

Cost of this serverless setup for a month use even for production scale $2 - $5

  • Lex : $0.38 (500 text requests)
  • Pinpoint : $0.06 (first 100 SMS FREE then $0.00645 per request)
  • Lambda : $0.20 (PER 1M REQUESTS, First 1M requests per month are free)
  • DynamoDB : $1.50 (First 25 GB per month is free + $1.25/million write + $0.25/million read)
  • Comprehend: $0.0001 (Upto 10M units)
  • Connect : $0.06 (For inbound calls in USA)
  • SNS : $0.50 (Per 1 million Amazon SNS, first 1 million is free)

Use cases

1. Greet the user with welcome message and HR menu options - Lambda Fulfilment

usecase1

2. User authentication and validation against identity database (DynamoDB in our case) - Lambda Validation

usecase2

3. Log in working hours, validate date and working hours

usecase3

> validate date

usecase3

> validate hours

usecase3

4. Calculate pay in a particular month, validate if existing user

usecase4

> validate if existing user

usecase4

5. Search common FAQs related to HR services, like Policy, Salary, Agent

usecase5

6. Sentiment Analysis of user interaction and connect to human support of negative sentiment found

usecase6

Integrations

1. Slack, FB, Kik, twilio

2. Amazon Pinpoint - two way SMS

3. Amazon Connect - connect center

4. Amazon Sumerian - talk to virtual host

5. Publish as Alexa Skill - Ask Alexa

Future enhancements and implementations:

  1. Integration with AR VR Amazon sumerian https://docs.sumerian.amazonaws.com/tutorials/create/beginner/dialogue-component/
  2. One-time passwords (OTP) to authenticate users https://aws.amazon.com/blogs/machine-learning/authenticate-users-with-one-time-passwords-in-amazon-lex-chatbots/
  3. Lex bot integration with Office 365 https://aws.amazon.com/blogs/machine-learning/schedule-an-appointment-in-office-365-using-an-amazon-lex-bot/
  4. Use languange transalation to make it multilingual bot https://aws.amazon.com/blogs/machine-learning/create-a-translator-chatbot-using-amazon-translate-and-amazon-lex/
  5. Document search using elasticseach https://aws.amazon.com/blogs/machine-learning/build-a-document-search-bot-using-amazon-lex-and-amazon-elasticsearch-service/
  6. Deploy on web https://aws.amazon.com/blogs/machine-learning/deploy-a-web-ui-for-your-chatbot/
  7. Build chatbot using Excel https://aws.amazon.com/blogs/machine-learning/build-an-amazon-lex-chatbot-with-microsoft-excel/
  8. Use voice authenitcation with Pindrop and Connect: Fraud Detection with Pindrop and Amazon Connect - https://aws.amazon.com/solutionspace/contact-center/amazon-connect-pindrop/

About

Build Amazon Lex chatbot to interact with Lambda, Amazon DynamoDB, Amazon Pinpoint, Amazon Comprehend, Amazon Connect, Twilio, Slack, Facebook, Kik and Alexa devices

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages