Intents

An intent in a conversation flow represents the specific goal or purpose a user aims to achieve when interacting with a AI-agent. For example, asking a question, making a request, or providing information. Intents help the AI-agent understand the user's inquiries and enable it to respond accurately to their queries.

For example, in a customer service AI-agent:

Intent: Check order status

User input: "Where is my order?:" or "Can you tell me the status of my recent purchase?".

AI-agent response: "Sure, please provide your order number."

Example 2:

Intent: Booking a flight

User input: "I want to book a flight to New York." or "Can you help me schedule a flight for next week?".

AI-agent response: "Of course, can you tell me your preferred travel dates?".

note

Yellow.ai DynamicNLP simplifies the process of model training by utilizing advanced technology called zero-shot learning. This eliminates the need for tedious and error-prone training processes.

Add Intents and Utterances

Utterances overview

Utterances are specific phrases or queries users use to express intentions. By mapping utterances to corresponding intents, the AI-agent can accurately understand and respond to user queries.

You can add intents (the purpose of the AI-agent users' responses) and utterances (the various ways users might phrase a specific intent, such as 'flight booking' or 'book flights' for the intent of booking a flight). You need to train the AI-agent to identify and understand these sentences to ensure accurate responses.

Best practices for creating intents and uttrances

To ensure accuracy and effectiveness of your AI-agent, adhere to these best practices while creating intents and utterances:

1. To effectively train the intents, ensure you create at least two distinct intents.
2. We recommend adding at least 15 to 20 diverse utterances for each intent.
3. Avoid using similar utterances both within a single flow and across multiple flows.
4. Ensure that all utterances in flows are unique to maintain clarity and accuracy.

Add Intents and Uttrances

To add intents and utterances, follow these steps:

1. Create a flow and add nodes based on your use case.

2. Go to Automation > Train > Intents > + Add new intent.

   

3. Enter the name of the intent and add the utterances. To add multiple utterances, click Add another utterance.

   

4. Click Add.

• This will add the Intent and Utterances to trigger a particular flow or to identify the user goal.

Search Intent

After creating an intent, you can use the search option to easily find the specific intent among multiple intents.

1. Enter the intent name in the search bar.

   

• This will display the intent that you are searching for.

Edit Intent

1. Hover over the specific intent to enable the Edit icon.

   

2. Click the Edit icon to modify the intent name, then click Save.

   

Delete Intent

To delete an intent, it should not be mapped to any flow. If an intent is mapped to a specific flow, it cannot be deleted.

1. Hover over the specific intent to enable the Delete icon.

   

2. Click Delete icon to remove the intent.

Add utterances manually

1. Click + Add to add the utterances to an intent.

   

Add auto-suggested utterances to the intent

Once you have added the intent, a set of auto-suggested utterances are generated. You can add your preferred utterances that are appropriate for your use case.

Add auto-generated utterances

To add auto-generated utterances, follow these steps:

1. In the right panel, under Utterance suggestions, click the Refresh button to generate a new set of utterances. Each click will provide a new set of suggestions.

   

2. Select the suggested utterances that are appropriate for your intent and click + Add selected.

   

• This will add the selected utterances to your intent.

2. You can also add specific suggestions by clicking the + icon next to each suggestion.

   

Edit suggested utterances

You can edit the suggested utterances based on your requirements before adding them:

1. Click the Edit icon next to each suggested utterance.

   

2. Make the necessary changes to the suggested utterance.

   

3. Click + iocn to add the edited utterance.

   

Train your AI-agent with intents and utterances

You need to train your AI-agent after adding intents and uttrances. If the AI-agent is not trained with the intents, it will not understand user queries and will display a fallback response. To improve user interaction, you can configure suggestions for the AI-agent to show the closest answers to the user query.

1. Add an intent and corresponding utterances as explained above.

2. Click Train intent. This will train the AI-agent with the provided intents and utterances.

   

3. If you have intents and uttrances in multiple languages, click on the Train intents drop-down.

   

4. Choose your preffered Feature type.

   

Option	Description
Sentence encoder	Converts input sentences or messages into numerical representations, making it easier for the AI-agent to understand and process user inputs.
Multilingual	If your AI-agent is designed to communicate in multiple languages, select this option to enable multilingual capabilities. This allows the AI-agent to communicate with users in different languages, to reach a wider audience.
Bahasa	Select this option if you want your AI-agent to communicate in Bahasa, the Indonesian language, allowing for more personalized and localized interactions.

note

If you have not chosen any of the options mentioned above, the Sentence encoder will be selected by default.

4. In Epochs, set the required number. Epoch refers to one complete iteration through the entire dataset during training. It helps the AI-agent improve its performance by learning from the data multiple times.

5. Click Train.

Trigger flows using intents

To trigger a flow based on a specific intent, follow these steps:

1. Navigate to the respective flow and click on the Start node.

2. Choose Intent as the trigger type.

   

3. Select the desired intent from the drop-down.

   

Test your intents

There are two ways in which you can test your intents- you can either test the flows in which they are set or you can test them in the Tools section. For more information, click here.

To test your AI-agent's confidence and intent identification:

1. Go to Tools.

   

2. Click Test your bot, enter the name of the intent that you have created, then click on Send icon.

   

• This will generate code to display the AI-agent's response to that intent.

In the example code below, the AI-agent confidently recognizes the phrase as part of the order intent, with a confidence score of 0.999.

{
  "text": "place order.",
  "intents": {
    "order": 0.999
  },
  "global_model": {},
  "intent": "order",
  "confidence": 0.999,
  "global_entities": [],
  "entities": {}
}

Stop or exit AI-agent conversations using intents

You can configure your AI-agent to stop or exit conversations using specific intents and utterances. Follow these steps to set up this functionality:

1. Add Intents and Utterances:

   • Create intents such as "stop the flow" or "exit the flow".
   • Add corresponding utterances that users might use to indicate they want to end the conversation.
   • Train the AI-agent on these intents and utterances.

2. Create a new flow

   • Go to the flow creation section.
   • Create a new flow and set the previously created intent as the start trigger.

3. Add a Text Message Node:

   • Within the flow, add a text message node with a message such as "Your chat has ended".

• By setting up this flow, whenever a user types any of the trained intents or utterances to stop or exit the conversation, the AI-agent immediately triggers this flow and ends the chat.

Resolve conflicts in intents and utterances

A AI-agent is trained with multiple intents and entities to improve its performance. If utterances are not classified correctly during training, it can lead to confusion. To resolve these clashes caused by unclear utterance classification, it is necessary to analyse the report.

Download utterance report

1. Go to Intents > Generate utterance report.

   

2. Two reports will be sent to your registered email address:

   • Utterances within intents/faqs: Identifies similar or highly diverse utterances that require editing within a specific intent/FAQ, including clashes caused by entity features.

   • Utterances across intents/FAQs: Identifies similar utterances across different flows and suggests modifying one of the similar utterances, taking into account clashes caused by entity features.

     

Resolve conflicts among utterances

Utterance reports sent to your email address help evaluate the quality of your AI-agent's utterances. They provide insights into the similarity of utterances within an intent and identify any common intents across different flows.

note

It is recommended to generate an utterance report after the initial Train setup and regularly at least once a month.

Conflicts within intents and FAQ's

This report identifies conflicts between two utterances if they have a high level of similarity. It compares utterances with the same intents.

If the similarity exceeds 50%, you need to visit the respective Intents page and either delete one of the similar utterances or rephrase the sentence.

If the similarity is less than 50%, ignore it.

Conflicts across intents and FAQ's

This is a comparison between utterances of different intents.

If the similarity is more than 50%, you must go to any of the Intents page and delete similar utterances or rephrase the sentence.

Best practices

The following are the recommended best practices to follow while naming intents and adding utteraces to them:

1. Best practices to follow while naming intents
2. Best practices to follow while adding utterances to the intents

Naming intents

For new AI-agents

AI-agents created after August 1, 2022.

1. Intent names must be at least 3 words long with unique words and no special characters.
   • Ensure intent names are descriptive to enhance clarity.
   • Avoid using generic names like "intent test one" or "FAQ number one".
   • Poor intent names can negatively impact NLP performance, leading to false positives and unnecessary AI-agent issues.
   • In the case of cloud, renaming intent names is possible.
2. The more descriptive the intent name, the better (add names with more than 3 words).
3. Avoid uncommon and business-specific abbreviations. Example: PO (purchase order ), GMV, and so on. Use the full forms and add synonyms if necessary. Few common abbreviations like UPI, EMI, and HR are acceptable.
4. Phrase the intent name as a verb followed by a noun. Example: get a premium receipt, pay renewal amount, fetch order status.
5. Keywords and sentences less than 3 words will fallback to the existing AI-agent model and will work as-is. These types of utterances will not go to the new model.
6. This model is applicable and works well for FAQs as well (since FAQs are descriptive and longer sentences)
7. Suggestions are automatically enabled for all new AI-agents by default. This default setting is crucial for improving the model and maximizing performance benefits.

For existing AI-agents

The following are a few important pointers for AI-agents created before August 1, 2022.

Cloud

1. Enable suggestions for AI-agents where they may not be enabled. This ensures that the model is used to the fullest.
2. Suggestions only show up for intents that are connected to the flows. Verify that unwanted flows are removed (or disconnected from intents).
3. If the intent name is camelCase (example: chatWithAgent) or has underscore/hyphens (example: chat_with_agent, chat-with-agent), use the edit option to rename these following the guidelines mentioned in the above section (for new AI-agents).
4. Ensure that there is no Small Talk in FAQs or Flows. If these are present, delete them. Platform small talk is enabled for all cloud AI-agents.

App (app.yellow.ai)

1. Enable suggestions for AI-agents.
   • To do this, in the App platform ensure that enableDidYouMean is set to true in app options in Function and in Tools > App Options > Prediction > Enable Suggestions.
   • If there’s an existing DidYouMean function in default:response, remove it.
2. Verify that the flow/journey DESCRIPTION is in line with the guidelines mentioned above.

   If these are not in line and are in camelCase or have special characters, change them by going to flow settings for that flow (you need not change the journey name, only the description can be changed).

3. Ensure that there is no Small Talk in FAQs/Flows. If these are present, delete them and enable platform small talk in Context Management and enable Small Talk.

---

Utterance & flows

DONTs ❌	DOs ✅
Do not add utterances in which the only variation is upper case or lower case	Do add at least 15-20 utterances per flow
Do not add utterances in which the only variation is Name, Date, City, and so on	Do ensure that there are an equal number of utterances in each flow
Do not create multiple flows that have a similar purpose	Do merge flows that are subsets of other flows
Do not overfit the model while training	Do use the didYouMean (suggestions) feature extensively
Do not add utterances if a flow will only be triggered through 'Trigger Journey'	Do minimize false positives
Avoid business-specific abbreviations	Do add abbreviations/short forms in the “synonyms” section
Do not add single words as utterances	Add complete sentences

• Do not add utterances in which the only variation is upper case/lower case

Wrong ❌

apply for leave

Apply for leave

apply For Leave

APPLY FOR LEAVE

Correct ✅

please apply for leave

can you please apply for leave?

how do I apply for leave?

procedure to apply for leave

This will make the model overfit and not learn the underlying sentence structure resulting in bad performance.

• Do not add utterances in which the only variation is Name, Date, City etc

Wrong ❌

apply for leave on 23rd Jan

apply for leave on 24th Jan

apply for leave on 5th Jan

Correct ✅

please apply for leave on 23rd Jan

can you please apply for leave tomorrow?

how do I apply for leave?

procedure to apply for leave

Wrong ❌

schedule meeting with John

Schedule meeting with Adam

Schedule meeting with Ram

Correct ✅

schedule meeting with John 

please block Adams calendar tomorrow for a meeting

sync up with Ram on January 3rd

• A few sentences (2–3) such as the ones listed below are acceptable but ensure that there are other utterances that show the different variations in sentence structure.

  • apply for leave tomorrow
  • apply for leave on 3rd

• Do add at least 15-20 utterances per flow

• The minimum number of utterances in each flow heavily depends on the complexity of the bot (number, type of flows and quality of the utterances)
• More utterances are always better especially when there are less than 10 flows.

• Do ensure that there are an equal number of utterances in each flow

Wrong ❌

Apply Leave flow - 50 Utterances

Leave Balance flow - 2 Utterances

Schedule Meeting flow - 30 Utterances

Correct ✅

Apply Leave flow - 50 Utterances 

Leave Balance flow - 50 Utterances 

Schedule Meeting flow - 50 Utterances

• Try to maintain a balance in the number of utterances per flow
• The NLP model is robust enough to handle small variations in the number of utterances (difference of 3-5 utterances)
• For smaller AI-agents (< 10 flows) maintaining balance is important to ensure good performance.

• Do not create multiple flows which have a similar purpose

Wrong ❌

Flow: talk-to-agent

Flow: transfer-to-live-chat

Flow: speak-to-agent

Correct ✅

Single Flow: talk-to-agent

• Having multiple flows which have similar utterances will confuse the model since there is a high amount of overlap.
• Merge all these flows into one single flow.

• Do merge flows that are subsets of other flows

Wrong ❌

Flow: apply-for-loan

Flow: apply-for-home-loan

Flow: apply-for-personal-loan

Flow: benefits-of-home-loan

Flow: benefits-of-automobile-loan

Correct ✅

Flow: apply-for-loan

    Steps: What kind of loan are you interested in?

        - Home, Personal, Automobile

Flow: benefits

    Steps: Which loan would you like to know more about?

        - Home, Personal, Automobile

• In the example above apply-for-home-loan is a subset or part of the apply-for-loan flow. This means that apply-for-home-loan will have utterances that are very similar to apply-for-loan

  - Example: “can you please help me apply for home loan?” , “Can you please apply for loan?”
  
  This will confuse the model during training 

• There are 2 steps to fix this:

  • Move all utterances to the parent flow in this case apply-for-loan
  • Create a step asking the user for additional details (in this case type of loan)
  • Another option is to setup entities. For example: type-of-loan - Personal, Home can be a type of entity within the flow.

• Do not add utterances if a flow will only be triggered through the Trigger Journey option

There may be some flows in the AI-agent that are meant to be triggered only from another flow (that is, these won't be triggered by the user’s input).

For example, Feedback Flow

For these flows do NOT add any user expressions or utterances. Adding utterances here will unnecessarily increase the complexity of the NLP Model.

• Do minimize false positives

  • When an incorrect flow is triggered with high confidence, it is considered a False Positive.
  • False Positives occur because of overfitting and spoil the customer or user experience.
  • These are minimized by following the best practices laid out in this document.
  • If there are a lot of false positives during training (even after checking for overfitting), try raising the "minConfidence" threshold.
  • The best strategy is to use didYouMean (Suggestions) feature and retrain the AI-agent periodically with the new data.

• Do use the didYouMean feature extensively

  • After ensuring that the model did not overfit (no False Positives) the next step is to enable the didYouMean feature
  • When the user’s input is not recognized by the model, the didYouMean feature elegantly handles the case as a fallback.
  • This is especially useful in the first few weeks after deploying a AI-agent in production, when accuracy may be low.
    • The self-learning capability allows the AI-agent to improve the confidence of different types of user expressions.

• Do add abbreviations/shortforms in the “synonyms” section

  • For short forms and abbreviations, add all the possible variations in the Synonyms section located under “Entities”.
  • The NLP pipeline will check for these abbreviations and replace them with the “full form” before passing them into the ML model, which will increase the accuracy.