Passer directement au contenu principal

We'd prefer it if you saw us at our best.

Pega.com is not optimized for Internet Explorer. For the optimal experience, please use:

Close Deprecation Notice

Video

Pega Intelligent Virtual Assistant: Unifying Chat and Email

Adding an intelligent bot doesn’t need to be another integration project. See how the Pega Intelligent Virtual Assistant unifies your channels and customer experiences by carrying a customer conversation, keeping context, and automating work across Facebook Messenger and email.


Transcript:

[Narrator] I'm gonna show an example where a user interacts with a company across several different channels. We'll start with Facebook Messenger. In this example, the customer is reaching out to find some details on a car. They'll load up Facebook Messenger, find their Dealerdirect car dealer and get started in the application. The buyer automatically recognizes the user and greets them by name and gives them some examples of what the bot can accomplish. It provides them a menu to choose from, but in this case, the user just types in text. The natural language processing is able to detect the user's intent and get them started working through a process to find the right car for them. You'll see several examples of how we're able to use Pega's user interface to help the user make suggestions and guide them along the path to the right vehicle. As we drive through this, we're able to narrow down more and more of what the right answer is for this particular user. We're also able to show them more detail by loading a web view, containing details of a specific automobile. But in this case, we really just wanna drive them to book a test drive.

We'll request their location and use the black box tools to send that into the application. We find the nearest dealer, present that to them user, and they can confirm that this is the one they'd like to look for. Again, we'll use natural language processing to find the right day for the test drive and let them select an available time at that dealership. We'll confirm this with the user and this go ahead and create both a scheduled appointment, but also generate a lead in the system for this particular customer. But we wanna have communication with them across other channels. So we'll ask to confirm a phone number in which to text them a reminder. You can see the text immediately pops up on the top here, confirming their appointment. Once the user confirms that there is nothing else we can wrap up this conversation. Now let's take a look at how a user would interact with our dealership via email. On the screen we see our email portal and this is what somebody in the back office and of the dealership would you use to triage the inbound emails? There are several emails here and we show who it was from what the subject was, but also some additional elements from our text analytics like sentiment and how the email was classified.

In the body of the email down below, you'll see highlighted pieces of text. These are things that have been identified by our AI as important relevant pieces of data that we might want to include in, in a case. We can also get additional details by going into the text analysis tab. This is more of an advanced feature, but can provide an additional layer of detail like filtering as well as the confidence with which we've identified the intent of this specific email. Once the users confirmed that this is indeed a request to find a specific vehicle, they'll click on that recommended case, and the details are populated in where available. So things like the city and the color can be automatically populated. The user can just simply click on the create a case, and that vehicle search can be initiated ultimately resulting in an outbound email to the customer.

They can then continue to work through their cases here. In this case, the customer didn't provide enough information so we're gonna send a templated an email response back to them requesting more information. But we really wanna be able to automate these processes. So let's show the design time experience and how that would work. Here we see our channel dashboard showing all the different ways in which a user can interact with our application. We're gonna drill into the email channel and we'll show how we can analyze attachments. We'll show the email addresses that we're listening on, but what we wanna do is update the routing. So currently it's getting routing all of those inbound emails to a work queue, and that's what that user was working from earlier. But instead, let's say we wanna go ahead and just automatically create a find a vehicle case whenever the confidence is above a certain threshold. Now when emails are sent in, rather than getting sent to the triage where somebody in the back office has to do the work, that case will automatically be created and an email will be sent back to the user. So let's send in an example here and what should happen is we should get a response back relatively quickly with all of the details we need from that email. So here we can see that we've received all of the details about the car we requested at a dealership nearby to us. This is all driven from a template based on the input from that user's email. So we've successfully and quickly automated this process.


Tags

Ressources associées

AI in Innovation

Exploitez la puissance de l'innovation par l'IA

Découvrez comment l'IA d'entreprise va contribuer à vos futurs résultats métier.

Why Pega

Pourquoi Pega ?

Notre puissant logiciel n'est pas notre seul atout.

Partager cette page Share via x Share via LinkedIn Copying...