Posts

Posts

  • Vikhil Khobragade
    Prediction and Recommendations in ODA45.0
    Topic posted July 29, 2019 by Vikhil KhobragadeRed Ribbon: 250+ Points, tagged Analytics, Bots, Conversational AI, Messaging, Notification, OMCE in Application Development > Digital Assistant and Mobile public
    Title:
    Prediction and Recommendations in ODA
    Summary:
    Implementation for prediction and recommendations
    Content:

    Hello,

    I'm looking for sample implementation for prediction and recommendations in Digital Assistant.

    e.g. For banking use case, when the existing customer log into the bot based on his customer profile or previous interaction to the bot.

    i would like to recommend him/her certain banking product i.e car loan or home loan.

    Can you please guide me how this can be implemented in ODA.

     

    Regards,

    Vikhil

  • Xena Arcabos
    Problem with the environment of oracle mobile cloud1
    Topic posted January 3, 2019 by Xena ArcabosGreen Ribbon: 100+ Points, tagged Bots, Conversational AI, Messaging, Mobile, OMCE in Application Development > Digital Assistant and Mobile public
    Title:
    Problem with the environment of oracle mobile cloud
    Summary:
    Getting no response from the Bots created in OMCE
    Content:

    Hi all,

    We are having problem with the environment of OMCE. We get no response from all the bots we created in the OMCE, even testing the bots in the environment itself. We checked the codes and its all good, we also checked the logs but no records listed. Can we have any explanation with this?
     

     

     

    Version:
    18.2.5
    Image:
  • Pawan Gupta
    Progressive updates with Digital Assistant..1
    Topic posted October 7, 2019 by Pawan GuptaRed Ribbon: 250+ Points, tagged Conversational AI, Virtual Assistant, VPA in Application Development > Digital Assistant and Mobile public
    Title:
    Progressive updates with Digital Assistant..
    Summary:
    Progressive updates with Digital Assistant..
    Content:

    Several times, requests or interaction with enterprise application takes time which could go beyond 30 seconds quite often, it may not be best for digital assistant to not respond with anything and have user wonder what's going on... many chatbot platform provide feature known as "Progressive update"... could you help me locate any blog or documentation if this feature exists in ODA.. I understand one way to do would be.. custom component can reply with no transition and keep the turn though question would be how to keep the handle on the REST call made to backend systems...

  • Ameur Baccoucha
    QnA componentAnswered105.0
    Topic posted September 19, 2018 by Ameur BaccouchaGold Medal: 3,500+ Points, tagged Bots, Conversational AI, Messaging, Mobile, OMCE, Tip in Application Development > Digital Assistant and Mobile public
    Title:
    QnA component
    Summary:
    i want to cretae a bot only using System.QnA component to make my bot reply to some question.
    Content:

    i want to build a sample of a Q&A bot. i created the bot and uploaded the csv file and every thing is ok but one i am using the bot tester i always get a response from the unresolved state. here my OBotML code: 

    metadata:
      platformVersion: 1.0
    main: true
    name: QnA_Bot
    context:
      variables:
    states:   
      qna:
        component: "System.QnA"
        transitions:
          actions:
            none: "unresolved"
          next: "qna"
      unresolved:
        component: "System.Output"
        properties:
          text: "Sorry, I did not find any match. Can you rephrase the question?"
        transitions:
          return: "unresolved"
      

    any idea please? 

    thank you 

    Ameur

  • dulce rodriguez
    question about System.OAuthAccountLink4
    Topic posted September 27, 2018 by dulce rodriguez, tagged Bots, Conversational AI, Messaging, Mobile, Virtual Assistant in Application Development > Digital Assistant and Mobile public
    Title:
    question about System.OAuthAccountLink
    Summary:
    component auth in bot with unwanted behavior
    Content:
    Hi!, 
     
    We are developing a chatbot using the Intelligent Bots product in the version 17.4.5 with Oracle Mobile Cloud Enterprise. Our chatbot will use authentication with LinkedIn like described in this tutorial:
     
    https://docs.oracle.com/en/cloud/paas/mobile-suite/use-chatbot/reference1.html#GUID-1C646528-1FE6-4248-AE7A-A9741B400521
     
    However, our users are experiencing a strange behavior when they authenticate.
     
    The following procedure describes the steps we have taken to include authentication functionality
     
    1. We created a chatbot using the "New bot" option
    2. We add a system component "System.OAuthAccountLink"
    In the field "authorizeURL" we put the url of the authentication provider with the following structure:
     
    PROVIDER_URL/authorization?Response_type=code&client_id=APPID&redirect_uri=URL_ORACLE.mobile.ocp.oraclecloud.com/connectors/v1/callback
     
    3. We save using the "Validate" option
     
    Apparently everything is fine, however when users run it they have the next behavior.
     
    a. Open the Bot.
    b. The bot goes to the LinkedIn site for authentication. 
    c. The user authenticate with valid data.
    d. Pressing the Allow button will download a response.txt file (unwanted behavior)
    e. The user has to manually close the browser to return to the chatbot (unwanted behavior)
     
    Could you help us avoid those unwanted behaviors?
    Version:
    17.4.5
    Image:
  • Silas Deblanc
    Recognizing names15.0
    Topic posted August 14, 2018 by Silas Deblanc, tagged Bots, Conversational AI in Application Development > Digital Assistant and Mobile public
    Title:
    Recognizing names
    Summary:
    Who is who- search by name function
    Content:

    Hi,

    I'm currently working on a chatbot project where I have to implement a "Who is who?" function. Here are some example questions to begin with:

    Search by function:
    Can you get me in contact with a warehouseman?
    - Help me find a project engineer.
    - I want to talk to a consultant.

    Search by department:
    -How can I contact the responsible of ICT?
    -I need to talk to someone from cardiology.

    Search by name:
    -Can i get the phonenumber from Guy Huyssmans?
    -I want the email of Deblanc Silas.
    -How can I get in contact with Maddy Desmet? 

    As seen above, I'm thinking to split up this "Who is who" into 3 different intents : Who_is_who-function , Who_is_who-department, Who_is_who-name. So I can make entities such as "function" and "department" (as value lists) and use their values to search through my database. 

    The problem now is that I have a database of over 3000 people so it's impossible to make a value list of all possible names. Is there any easy way of recognizing and extracting names out of a sentence in Oracle Chatbot? Or do you have an alternative solution? 

    P.S: The names are mainly dutch names.

    Thanks in advance,

    Deblanc Silas

  • Xena Arcabos
    Replicating Explicit invocation prompt3
    Topic posted June 20, 2019 by Xena ArcabosGreen Ribbon: 100+ Points, tagged Bots, Conversational AI, OMCE, Virtual Assistant in Application Development > Digital Assistant and Mobile public
    Title:
    Replicating Explicit invocation prompt
    Content:

    Hey Guys!

    Is there any other way we can replicate the Explicit invocation prompt of the DA like when a user inputs a phrase into the digital assistant, the digital assistant determines how to route the conversation then a prompt will pop up asking the user if he wants to continue going to that skill. But we only want it within a skill where in switching from intent-to-another intent.

     We are using the action-transition textReceived: "intent" but we want to have a prompt asking the user if he wants to continue going to another intent or remain to the current flow.

     

    thanks,

    Xena

    Version:
    19.1.3-1903140545
  • Tom Smits
    Resolve Composite Bag Variabele6
    Topic posted November 16, 2018 by Tom SmitsGreen Ribbon: 100+ Points, tagged Conversational AI, Mobile, Virtual Assistant in Application Development > Digital Assistant and Mobile public
    Title:
    Resolve Composite Bag Variabele
    Summary:
    Using the commonResponse component, a composite bag is only valued when the first response item is answered
    Content:

    Hi all,

    Currently, I'm working on a Digital Assistant that creates an appointment request for potential leads. I created a composite bag with the entities date, time, location and employee. However, when i ask something like this: "can i meet with "John Doe" tomorrow at 12pm in Amsterdam?". The bot asks for a date. When i respond with "tomorrow at 12pm in Amsterdam" it fills the appointment composite bag with all the matched entities.

    Why didn't this happen when i was asking the first question? I've provided the code below.

    Kind Regards,

    Tom

     

     

    Version:
    Version 18.3.2_1808150141
    Image:
    Code Snippet:
  • Sharath Chandra Gavini
    Resolve Conflicting Entities
    Topic posted June 29, 2019 by Sharath Chandra GaviniBronze Medal: 1,250+ Points, tagged Bots, Conversational AI in Application Development > Digital Assistant and Mobile public
    Title:
    Resolve Conflicting Entities
    Summary:
    Prompt the user to choose one of the entities when a single utterance resolves multiple values
    Content:

    Hi developers,
    I have a use case where a single utterance resolves multiple entites which I would explain soon. I want to know how the user can be shown a prompt without using ResolveEntites component ( I am not using an nlpresult variable, using it needs a lot of change in the conversation flow and the bot is in production). I used System.MatchEntity and as we all know it resolves only one Entity value, it doesn't fit in this place. I want to display the user a list of choices with multiple values without using the nlpresult variable (like I mentioned earlier, intent processing is not done at the stage). I tried using System.Text and even it resolves a single Entity value.

    My use case: 
    I am resolving country name based on the user input and People's republic of China is an utterance for China and Republic of China is an utterance of Taiwan. 
    When the user types People's republic of China, the bot resolves Taiwan implicitly. I want to display the user a prompt to choose between China and Taiwan and it's the best if People's republic of China resolves China only. My training model is Machine Learning model. 
    Please let me know if there is any possibility of achieving this without the usage of an nlpresult variable.
    I am insisting on not using an nlpresult variable as there are multiple intents and it would change the entire flow to incorporate this small change.
    Thanks in advance,
    Sharath Chandra Gavini.

  • Ameur Baccoucha
    Sending FilesAnswered9
    Topic posted October 30, 2018 by Ameur BaccouchaGold Medal: 3,500+ Points, tagged Bots, Conversational AI, Messaging, Virtual Assistant in Application Development > Digital Assistant and Mobile public
    Title:
    Sending Files
    Summary:
    i wanna that bot users send me files (image, audio, PDF ...) how can i get them ?
    Content:

    Hi everybody,
    I am asking how to receive files sent by the user with FB in different formats such as audio, image, PDF .....
    also if there is a documentation on how to integrate speech recognition.
    thanks

    Version:
    ODA 18.3.2_1808150141