/ChatbotWithLexandLambda

A conversational interface that can answer common questions or perform specific tasks.

Primary LanguagePython

CHATBOT WITH LEX AND LAMBDA

PROMPT

Build a chatbot using Amazon Lex and AWS Lambda. Create a conversational interface that can answer common questions or perform specific tasks. Integrate the chatbot with a messaging platform like Facebook Messenger or Slack.

ARCHITECTURE DIAGRAM

STEPS

  1. Navigate to Lex

  2. Create a bot

    • Create blank bot

    • Enter Bot name as "SupportBot

    • Under IAM permissions

      • Seclect "Create a role with Amazon Lex"

    • Under Children’s Online Privacy Protection Act (COPPA)

      • Choose No

    • Click Next

    • Select Language

    • Choose Voice Interaction eg English (US)

    • Click Done

    • Intent: New Intent

      • Intent Details
        • Enter "SupportIntent" as Intent name
      • Add Sample utterances (utterances, are what the user can use to start interaction with the chatbot)
        • Type an Utterance eg hello, hi
        • Click Add utterance
      • Toggle "Confirmation" to active
      • Same with Closing response
      • Under Confirmation,
        • enter the following in the confirmation prompt: "Can I go ahead with your request?"
        • enter the follwing in the decline prompt: "Okay, your request will not be submitted."
      • Under Closing response
        • enter the message as: (response chatbot will give when the user starts interacting with their utterances) Thanks!
        • (Optional) can add variations to spice things up
          • Click "More responses under Variations
            • The closing message editor will pop up
            • Click on the variations dropdown and enter "Thank you!"
            • Click Update responses
      • Save Intent
      • Slots are placeholders for the user input

      • Navigate back to intents list

      • In the Amazon Lex menu, click "slot types" under OrderPizza -Click Add slot type dropdown - Click "add blank slot type"

        • Specify Slot type name

          • Enter "PizzaType" as Slot type name
          • Click Add

        • Under Slot value resolution

          • Select "Restrict slot values"

        • Under slot type values

          • Under value, enter "Italian"
          • for its corresponding values, enter italian, italy,Italy
          • click tab or ; after each value, for a new one
          • You can add more values like "Mexican", with corresponding values like "mexico", "mexican" etc

        • Save Slot type

        • Navigate back to Slot types

          • Create two more: PizzaCrust and Appetizer

      • Navigate back to Intent: OrderPizza

      • Slots

      • Add slot

        1. Name: Name Slot type: AMAZON.firstname Prompts: Hello! May I know your first name

      • Click add

      • add more slots:

        • Pizza Type
          • name: PizzaType
          • slot type: pizzatype
          • prompts: {Name}, What pizza would you like to order?
        • Add custom value by putting the slot name in {} eg {Name}

      • scroll up to Conversational flow to see how the conversation will go

      • Save intent

      • Improve UI by adding values

        • Choose DeliveryTime option
          • click advanced options
          • scroll down to slot prompts
          • click bot elicits information DROPDOWN
          • cLICK more prompt options
          • toggle preview -click add -add card group
            • update card

              • enter title as "Available times" as placeholder

              • click buttons fropdown click add button - button 1 title : 20:00 - button 1 value : 8pm

                  - button 2 title : 20:30
  - button 2 value : 8:30pm
            • click "update prompts"
            • click update slot

      • Save intent

      • Click Build

      • Test

  3. Navigate to Slack

    • Create an workspace
      • https://slack.com/
      • Click Try for free
      • Enter your mail
      • Create workspace
      • Step 1: Enter the name of your company or team
      • Step 2: Enter your name
      • Step 3: Skip "Who else is on the SupportTees team"
      • Step 4: Enter "SupportBot for Teesside Uni" as what ypur team is working on

  4. Navigate to slack api

    • https://api.slack.com/
    • Click "Your app" at the top right corner > Create ypur first app
    • Select "From Scratch"
    • Name app & choose workspace
      • Enter "SupportTeesBot" as app name
      • Select "SupportTees" as workspace
      • Create App
    • Go to Setting > Basic Information
    • Scroll down to App Credentials
    • Copy the following credentials from slack api and paste in Lex console: add channel
      • Client ID
      • Client Secret
      • Verification token
    • In the Lex console, click "Add"

  5. Configure Integrativity & Shortcuts

    • Go to Features > Integrativity & Shortcuts
    • Toggle on
    • Paste "https://slack.com" as the Request URL (temporarily)
    • Save changes

  6. Set up the OAuth Endpoint

    • Navigate back to the Lex console
    • Click the newly created channel
    • Scroll down to OAuth endpoint, and copy it
    • Navigate back to slack api
    • Go to Features > OAuth & Permissions
    • Add new redirect URL
      • Paste
      • Click Add
      • Click Save URLS
    • Scroll down to Scopes
      • Under Bot Token Scopes, click Add OAuth scope
      • Add:
        • chat:write
        • team:read
    • Navigate back to Lex and copy Endpont url
    • Backl to slack api > features > interactivity & shortcuts -Replace the temporary url with the endpoint url
      • Save changes

  7. Events & Subscriptions

    • Toggle Enable events

  8. Channel Integrations

    • Select "Slack" as platform
    • Leave IAM role as is, assuming its the Amazon Lex role
    • Under Integration configuration
      • Enter "SupportTees" as Name
      • Enter "Customer support for prospective TU students" as the description
      • Select "TestBotAlias as Alias
      • Select "English (US) as Language (should be the same as what was chosen when creating the bot)

RESULTS