Global Intents

Global Intents

Overview                                               

Global Intents Steps add the benefit of Natural Language Understanding to your bots and conversations.

Global Intents, allows the Bot to use the intent recognition mechanism, identify what intent the customer expresses and use it to progress through the conversation path.

Intents can be used to handle a wide range of customer requests, from simple questions to complex questions. When an intent is matched,  the Bot will follow the path configured for that specific intent.

The biggest advantage of using Global Intents instead of a rules-based Listening Step, where customers must return pre-determined options and follow specific syntax or commands to express their intent, is that the customer can formulate their request in a variety of ways, and Global Intents will still be able to understand the intent behind a message and and route the conversation where you want. 

When configuring a Global Intent step you can choose what 'Engine' you want your Bot to use. You can select either WebioFlow or DialogFlow.
Info
Global Intents & WebioFlow are currently released to selected clients, if you want to get involved get in touch with us at info@webio.com

WebioFlow

When you select Webioflow, you access pre-made Intents from our default AI Engine Intents pool.
Alert
If you have a bespoke model you will access your own model Intents pool. 


Our default model is finely tuned to the world of Credit & Collection and can easily identify the most common intentions of your customers, such as the intention to pay or express a complaint to make the bot take the path you have configured for the intent that is being matched.


The WebioFlow engine will dramatically improve the Bot's ability to understand the intent behind a customer response and respond with a relevant message/question that you have already configured, in this way you will further automate and improve the triage and payment collection processes.


                                           

Warning
Please note that "Late Payment" Intent (used in the example above) is not yet available on our new Web App. Contact support@webio.com if you want to know more.

How to enable and use WebioFlow AI Intents?

When you're on the bot's build screen.
  1. Open the Action category, on the left side of the screen, and drag and drop on the bot building screen, a Global Intent step.
  2. Select WebioFlow.


  1. Compose your message. 

  1. Open the A.I. category, select the intent you need (check below for a full list of intents), and drag and drop it into the "Drag & Drop Intent" box.

  1. Save your configuration.

Info
You can add as many intents as you want in the Global Intents steps.
  1. If no intent is matched to the customer's message, the Bot will follow the default default fallback (at the bottom of the tab) and the path you've configured for it.


  1. You can also configure a manual fallback. This one will take precedence over the previous one. However, if the A.I. does not identify any of the intent mapped intents, the Bot will follow the automatic fallback to the bottom of the Step. It is recommended to configure both and connect them to the same default step.


  1. Configure a follow-up step for each of the intents used in the Global Intent Step.


What WebioFlow can understands?

Info
Please check this link to learn more about WebioFlow Intents abilities.

WebioFlow AI Embedded Entities

When you use Webioflow, you also access as well the pool of Entities. In this way you can embed entities to a specific intent.



Embedded Entities can be used as a follow-up to any of the intents. WebioFlow Entities is able to analyse the message provided by the customer and understand if any of the following entities have been provided, regardless of the format or order that are returned:
  1. Phone number 
  2. Date  
  3. Name 
  4. Number
  5. Post Code
  6. Amount of Money & Currency
  7. Custom Entity - Please get in contact with support@webio.com if you want to enable this feature.

How to use Embedded Entities

Alert
Please note that Embedded Entities can only be used in conjunction with the intents configured on a Global Intents step. This will use the AI model that is available to your organisation.

If you are looking for information regarding standalone Entity Gathering, please follow this link. link. Please note that Entity Gathering step doesn't use bespoke models.   
Once you've added an Intent to the Global Intent Step:
  1. Click on ADD ENTITY

  1. Drag and drop the desired entity into the Entities Drag & Drop box

  1. Select in which Field name you want to store customer responses. You can add multiple Entities for the same Intent. (E.G. Name and date of birth) 


  1. Compose your message. Ensure you ask for the Entities your are looking for.


  1. Finish the configuration of your Global Intents step.
                                                                                    
  1. Hit SAVE. 
Your Global Intents Step is now ready. You will need to configure the rest of the Bot's steps and create the follow-up path for each of the intents and entities that could be identified by your Global Intents step.

E.G.



Info
When using Entities within a Global Intents step, you can use a failure fallback to configure the path that your bot should take in case one or more entities are not returned or detected in the customer's response message.

DialogFlow

When you select Google DialogFlow as your Global Intent Step engine, you access all intents from your Google DialogFlow agent.

If you want to use DialogFlow:
  1. Open the "Action" category and drag and drop a Global Intents step onto the building screen.
  2. Select DialogFlow. 
  3. In the Bots field, select the desired DialogFlow agent. (Make sure you've integrated your DialogFlow agent from the Organizations - Webio - 

WebioFlow - Some examples

  1. How to build an ID&V Bot using Global Intents and Logical Branch steps

In the example below, we used a Global Intents step powered by WebioFlow engine (Step 2) and provided the customer with two options:

1. Checking their account details. 
2. Performing a payment.

If the customer returns an intent identified by the "Account Enquiry" intent, the Bot will follow up with a message asking them to confirm their name and date of birth.

If the customer's response includes a valid name and date of birth, these two entities will be saved in the custom fields that we called "full name" and "DOB" respectively (Step 2).

We then have used two logical branches (Step 5 and 6) to to verify that the customer had returned values that matched our records. 





When the customer, in response to the Global Intents (Step 2) returns the entities we are looking for, the Logical Branches will check if the details provided match or not the records imported from our campaign file (or your internal system for Webhooks integrations).

The Bot will then take the configured path depending on whether the conditions in the two Logical Branches (Step 5 and 6) are met or not.

 

 

  1. How to capture an amount of money (Promise To Pay) using Global Intents 

Using Global Intents, we first asked the customer to confirm whether they can make a payment or not. (Step 1)

If the customer proposes a potential payment, this intent will be identified by the "Promise" intent. (Step 1)

When the 'Promise' intent is identified, the Bot will then proceed to collect the WebioFlow entity 'Amount of Money' in order to collect the amount of money the customer is willing to pay.

WebioFlow Entity analyzes the structure of the customer's messages in search of the entities you have selected. It doesn't matter if the customer returns the requested entity in a non-specific format

Webio Flow Entities can understand that the amount of money the client is willing to pay is £100, no matter if the client replies:

"100£"
"I'm gonna pay 100£"
"I  think I can pay 100£"
Thanks to a Logical Branch (Step 2),  we have then set a threshold of 250£ to verify if the amount the customer is willing to repay acceptable or not.



  1. How To Build A Quick Survey Bot Using Global Intents 




    • Related Articles

    • How to integrate a DialogFlow agent

      Overview Dialogflow agents use natural language understanding to return intents. When used as part of your Webio Conversational Middleware solution they allow your bots to manage conversations with natural fluency, allowing a conversational tone that ...

    • WebioFlow Intents

      Overview WebioFlow pre-trained intents provide your Bot with the ability to identify your customers' intentions, in particular requests related to the Credit & Collection environment (Payments, payment promises, etc.) and positive and negative ...

    • Step types

      Overview Bots are created as a flow of steps. The Bot's automatic conversation will follow the flow according to the rules and the path you have configured. When you create a new step, you need to choose what type of step it is, configure it, and ...

    • Action Panels

      Overview When you open a conversation in the Conversations screen, the Action Panels (on the top right side of the screen), provide you with tools and information to maximize your team's productivity and streamline the way the conversation is ...

    • How to build an Inbound Bot

      Overview An inbound bot collects the channel ID from which the contact's message is sent, and uses it to send messages back. Once the customer's inbound message is received, to continue the conversation, you must necessarily use a Listening, or a ...