Related questions

Lex selects one intent for any one-word utterance that doesn't match a valid utterance

2018-10-11 19:37 Jay Gerring imported from Stackoverflow

For a few months, my team has been working on a Lex-based chatbot that's attached to a marketing website. The bot has a total of 32 intents that handle over 300 utterances right now. In an effort to reduce utterances that invoke error handling responses (and to make the bot more useful and friendly), we've implemented 3 intents in the following order over time:

1) a "Thank You" intent: responds with "you're welcome"-type messages when a user types some variant of "thank you".

2) a "No Answer Available" intent to handle known utterances (from our usability and internal user testing) that the bot isn't currently designed to answer, but that we want to refer the user to a human chat agent (e.g. "I don't have an answer for that, but type 'transfer' to talk to one of our human agents who can help."). This intent currently has the largest set of utterances in the bot.

3) a "No Response" intent that "eats" a small number of utterances from the user that don't really require a response and would probably go to error handling other wise and make the bot look dumb, e.g. "you're welcome", "ok, got it", "cool", "k" etc. IOW, it doesn't return any response for these to the user.

Until we implemented #3 (maybe coincidental?), if a user typed 1 or more nonsense words to the bot, it properly triggered error handling. After implementing #3, a single, alpha-only word of apparently any length that doesn't match a valid utterance is now triggering the "Thank you" intent.

Screenshot: bot says "you're welcome" to a nonsense word

Note that 2 nonsense words triggers the expected error handling messages. Also, single strings that contain symbols or numbers also trigger error handling as expected. Interestingly (and unexpectedly), text-based emoticons typed in the usual direction, e.g. :-) or :) or ;-) also trigger the "thank you" intent, but if you type them in the reverse direction, e.g. (-: they trigger error handling.

Additionally, 2 other intents handle a number of single-word utterances:

1) A "signoff" intent that handles words like: bye, goodbye, later, exit, quit, etc.

2) A "hello" intent that responds to: hi, hello, hey, what's up, what's happening, etc.

So the "Thank you" intent has neither the largest number of utterances in the bot nor the largest number of 1-word utterances (just "thanks", actually) both which I have seen in various forums as reasons why a Lex intent would attract incorrect utterances.

I've tried searching for answers on this phrasing the question any way I can think of and haven't found anything.

So, though it's not a product show-stopper, it's making me a little crazy, and I thought I'd post here to see if anyone has run into anything like this in their Amazon Lex journeys and had any ideas.

Thanks!

See also questions close to this topic

  • Is it possible to specify a second file to build requirements from using AWS Serverless Build?

    Let's say I have a template.yaml which has the following resource:

    Resources:
  TesterFunction:
    Type: AWS::Serverless::Function 
    Properties:
      CodeUri: test/
      Handler: app.tester
      Runtime: python3.7
      Events:
        TestFunc:
          Type: Api 
          Properties:
            Path: /tester 
            Method: get

    When I run sam build it automatically looks for requirements.txt in CodeUri path. Is there a way to specify a second requirements file to build from e.g env-requirements.txt?

  • Matching certificate for VPC-Link HTTPS endpoint in API Gateway

    I would like to change my API from HTTP to HTTPs in API Gateway, but get the following error: 

    Endpoint request headers: {x-amzn-apigateway-api-id=ghqsfz8w25, 
Accept=application/json, User-Agent=AmazonAPIGateway_xxxx, 
Host=xxx-nlb-ecs-service-xxx-xxxcea41.elb.eu-west-1.amazonaws.com, 
X-Amzn-Trace-Id=Root=1-xxx-xxx}
Endpoint request body after transformations: 
Sending request to https://xxx-nlb-ecs-service-xxx-xxxcea41.elb.eu-west-1.amazonaws.com
Execution failed due to configuration error: 
Host name 'vpce-xxx007-xxxb1h.vpce-svc-xxx318ce.eu-west-1.vpce.amazonaws.com' does 
not match the certificate subject provided by the peer (CN=*.domain.eu)
Method completed with status: 500

    With the following configuration:

    • A wildcard certificate *.domain.eu is used
    • The wildcard certificate is assigned to the NLB 443 listener
    • The NLB is integrated with vpce-svc-xxx318ce endpoint in VPC Endpoint Services (AWS PrivateLink)
    • The wildcard certificate is assigned to the custom domain in API-GW
    • The custom domain includes mappings to different APIs (like: api.domain.eu/api1v1, api.domain.eu/api2v1)

    The settings of the API method are

    With HTTP this configuration was working. Now it seems that I need to associate the VPC endpoint with the wildcard certificate. It is my first experience with AWS and I am not a network expert. Any ideas how I can fix this issue.

    Thanks in advance!

  • Removing specific log events from CloudWatchLogs

    I looked through the documentation of CloudWatch Logs, but I could not find any method to manually remove partial data from a log stream.

    The only way to remove data is to:

    • Remove the log stream, but that will remove all data (not just the data we want to remove)
    • Set the retention policy to a smaller time window, so that the log events older than X time will be deleted. This also removes more than we'd like to remove.

    Does anyone have experience with removing specific log records?

  • How to add a IBM Watson Assistant chatbot on to a webpage?

    I am trying to add my chatbot to a webpage but cannot use the IBM cli for this due to me not having admin rights on my system for install. I have eclipse as an IDE and cannot find good documentation for the above mentioned integration.

  • Dialogflow "Training" menu is empty always

    I am working on building a dialogflow agent. The "Training" menu used to provide me the list of utterances which didn't match any of the intents. I could check this menu and assign to the right intent and approve it to get trained again.

    Now, I could see only an empty page. The list of utterances are not appearing. None of the configuration got changed and I checked settings as well to see if logging is enabled and it is enabled as expected.

    I have read through all the docs and compared with my other project settings too and nothing seems to be missed out.

    The "Training" menu remains empty. Would be great if I can get some help on this.

  • Passing the value of JSON Object from HTTP Request to another JSON Object

    I'm working on simple Chatbot in Line Messenger. I found some API which I want to use in the following way:

    User sends specific key word like !send data. Then I send HTTP Request using request module. I'm getting the response and I output it in the console just to see if everything is correct. I'm parsing the repsonse with JSON.parse(body) and I'm able to access its specific values with 

    result = JSON.parse(body);
console.log(result.value.text);

    Now I want to send result.value.text as a reply mesaage to the Line client. Message has form 

    message = {
  type: "text",
  text: "someString"
}

    But how can I pass the part of HTTP response to this object? Trying something like this

    LINEResponse = {
  type: "text",
  text: result.value.text
};

    leads to 

    UnhandledPromiseRejectionWarning: Error: Request failed with status code 400

    I'm not 100% sure if its even correct approach implement this. I would appreciate any help

  • How can I return hyperlinked text in a Lexresponse?

    So I am building a lex chatbot and I am trying to return a response with hyperlinked text. I have the chatbot sitting on a front end but I cant seem to find a way to return responses with hyperlinks. Heres what I have so far

    https://imgur.com/N6Bp2fX

    https://imgur.com/zbnUsrH

    Now Ive read that the responses from lex are formatted to where the chatbot is sitting. For example, in the chatbot test window on the Amazon site, returning hyperlinks is impossible, but skype automatically hyperlinks URLs. But I have mine sitting on a browser but I still cant get a hyperlinked response in the bot.

    Would love if anyone could help me out! Thanks in advance!

  • Can we get multiple lines long response from user with Amazon Connect and Lex

    I am working on Amazon Connect with Lex and Lambda integration. I have requirement to get user feedback. User can give long feedback. Right now, Lex is taking only 2 lines of user's response. Is there any way to increase that limit? If not, how can I achieve it using Amazon Connect?

    Thanks, Gan

  • How to connect Lambda to an API?

    I am an intern working on a chatbot that uses lambda functions and I am trying to connect the chatbot to our companies internal API.

    It seems like Amazons API Gateway is the only option and that requires you to build your own API.

    If anyone could guide me in the right direction, even connecting to any simple existing APIs, that would help alot! Thank you