Configuring AI Launchpad

 

AI Launchpad eases the enablement process of any Talkdesk AI product by providing easy-to-use configuration capabilities.

By using AI Launchpad, you can enable Talkdesk AI products in a self-serve manner.

Screenshot 2024-08-01 at 16.20.37 copy.png

When you first access the AI Launchpad app, you will be presented with the “My area” main page. On this page, you will see the access points (onboarding cards) to help you configure the different AI flows for each of the products.

Below the onboarding cards, you will also see all the products that you already have installed. 

To understand how to configure AI Launchpad, please read through the instructions below. 



My Area

Enabling Voice Transcription

Products that need voice transcription such as Copilot, and Customer Experience Analytics, should be first configured by opening the Voice Transcription card. Then, the configuration should happen as follows:

Screenshot 2024-08-01 at 17.24.48.png

1. Select the product that you wish to configure [1] and click Next [2].

Screenshot 2024-08-01 at 17.27.21.png

2. Select the language(s) that you wish to use in your account by clicking on the respective card(s) [3]. The languages presented in the tiles are the most used languages. Click to select one or more of these languages. These correspond to the ones that you wish to use. Later, you can associate them with “segments”, i.e., ring groups and/or agents.

3. If the language that you wish to use is not represented in any of the cards, select it from the “More options” dropdown menu [4]

4. Select a language for the default AI flow [5]. The default AI flow will consider the first language selected (or a different one if explicitly selected) as the default one. Selecting this default language is not mandatory, but it serves the purpose of having a rule that is applied to any conversation which is not covered by any of the more comprehensive rules, such as those defined by ring groups or agents. This language rule will automatically encompass all agents and/or ring groups, as the following window will show. 

5. Click Next [6].

Screenshot 2024-08-01 at 17.57.28.png

6. Add the “segments” to each of your AI flows by selecting the Agents or Ring Groups option from the dropdown, along with the respective agents or ring groups [7]. You can also choose the segments within the ring group or agent listings [8].

  • Note: Ring group and Agent information will be used to select which flow to apply to any given interaction. The priority of the flows will be determined as follows:
      • First look into any flows that contain the call Agent configured at Agent level.
      • If no Agent-level flows exist, look for any that contain the ring group associated with the call.
      • Lastly, use the default flow, in case it exists.

7. Select the call direction from the following options: Inbound, Outbound, or Inbound & Outbound [9].

8. Add as many segments to your AI flow as desired by clicking Add [10] and selecting the options as described in 6 and 7.

9. Click Next [11].

Screenshot 2024-08-01 at 17.59.21.png

10. Review the configurations of the AI flow(s) that are going to be created. Click All Set [12] to end the configuration process and publish the AI flow(s).

Editing Users

Note: For this example, we will use Talkdesk Copilot.

Screenshot 2024-08-05 at 16.31.15.png

In order to further edit the users that can access each Talkdesk product that was configured in the earlier step, Admins must click on the icon at the bottom-left of the screen [1]. Locate "Copilot" and click the wheel icon [2] that appears as you hover over it.  

Screenshot 2024-08-05 at 15.32.52.png

A menu will appear, where you will be able to define the roles and/or individuals that have access to using the product, via the "Users" [3] tab. Tick the boxes with roles/users you wish to enable and then press the Save changes [4] button.

Screenshot 2024-08-06 at 09.24.44.png

A confirmation message will appear, prompting you to Submit [5] your changes.

Screenshot 2024-08-06 at 09.25.02.png

You can also utilize the "Auto Install" tab [6], to select the roles that will be provided access. Follow the same process as above, by pressing the Save changes button and then Submit, to apply your selection and provide access to those roles.

 

Adding AI Models to your Speech-to-Text Configurations

Once the voice transcription has been configured in the languages of choice for Copilot or Customer Experience Analytics, AI Models can be added to each AI flow. 

From the AI Launchpad main page, you can enrich your flow with an AI Model by: Screenshot 2024-08-01 at 16.20.37.png

  • Clicking on the AI Models card on AI Launchpad main page.

Screenshot 2024-08-01 at 16.20.53 copy.png

  • Then you must select the AI product that you wish to configure, and clicking Next.

Screenshot 2024-08-01 at 16.17.47.png

  • You can also locate the flow to which you will be adding the AI Model [1] and selecting the “Enrich with AI Models“ button [2] relevant to the flow that you wish to modify. This button will only appear if the flow still hasn’t got an AI Model associated with it.

In any case, you will be directed to the page below where you will be able to perform the following configurations: 

Screenshot 2024-08-01 at 18.15.57.png

  • Select the AI flow to which you want to apply the AI Model [3].
  • When doing so, you will get a listing of all the AI Model possibilities regarding your language. Select the AI Model that you wish to apply to your AI flow [4].
  • If you do not wish to use any of the available AI Models, you can create a new AI Model by clicking on the Create intent model option [5]. Please see the AI Resources section for more information on how to do it.
  • The “Preview” section will show you a summary of your selection [6].
  • Click Next [7].

Screenshot 2024-08-01 at 18.17.39.png

  • Choose a name for the AI flow [8] and, if necessary, add a description [9].
  • Click All set to confirm the selection of your AI model [10].

 

AI Resources

On this page, you can see and manage your AI Models. 

There are two types of AI Models: 

Intent Models

An intent model is a general-purpose, trainable, intent detection model. It’s a classifier that accepts text as input and outputs one or more pre-configured classes. An intent model can be attached to a variety of product flows and can be shared across them.

Conversational Models

Conversational models are conversational intent models that support conversational use cases, such as the ones for Autopilot and Copilot’s Smart Scripts.

From the several types of Conversational models available, we are currently offering support for Conversational models which can be applied to our portfolio of products as seen below. Please be aware that a Conversational model can be created in AI Launchpad but will not be ready to be used with any of your AI flows until an automation is published within that model. This can be done in Automation Designer and is a mandatory step for your Conversational models to be listed in AI Launchpad during the enrichment stage of any AI flow.

Conversational Flows: 

  • Autopilot - Customer-facing, voice, digital.
  • Smart Scripts - Agent-facing, digital.
  • Feedback Surveys - Voice and digital.

 

Creating an AI Model

Screenshot 2024-08-01 at 18.18.56.png

1. Open the AI Resources tab, and then click on Create new [1].

Screenshot 2024-08-01 at 18.21.49.png

2. A side drawer will open where you will be able to customize your new AI Model. Fill in the options as described below:

  • “AI Model type” [2]: Select Intent Model or Conversational model.
  • “Name” [3]: Add a name to your AI model.
  • “Description” [4]: If necessary, add a description to our AI model.
  • “Service Provider” [5]Select a provider and then, a language.

3. Finish the creation process by clicking Create AI model [6].

Be aware that when creating AI Models from the AI Resources page, the created model will not be associated with any of your AI flows and will, therefore, not be in use until you add it to one of your flows. Read through the Adding AI Models to your Speech-to-Text Configurations section to learn how to do it.

It is also possible to create an AI Model in the edit screen of a specific AI flow:

Screenshot 2024-08-01 at 18.25.06.png

4. Click on the more options icon [...] [1]of the flow for which you’d like to create a new AI Model, followed by Edit [2].

Screenshot 2024-08-01 at 18.30.18.png

5. Then, on the Edit screen, located the Intent Model tile and click on the Add AI model type button [3].

6. After this step, you will be taken to the same page mentioned in point 1; The configuration steps will also be the same as the ones described there.

If you’re creating the AI Model directly in the edit screen of a specific AI flow,  this AI Model will automatically be selected by default so that when you save your flow, the model will start being used within the context of that flow. Please note that this model will also be made available for use in other flows, but to enable its use in those additional flows, you need to edit each and select the proper model.

Tiers

Screenshot 2024-08-01 at 10.18.38.png

All products installed in AI Launchpad have a corresponding tier. This tier is a reference so that you know to which capabilities you are entitled.

In AI Launchpad, there are 3 different tiers, represented by tags - “Starter”,Standard”, and “Premium”.

All Articles ""
Please sign in to submit a request.