Chatbots

Voice Digital Assistants

Every day we are more accustomed to use digital assistants such as Google Assistant, Alexa, Siri and Cortana, among others, who accompany us and help us with our daily tasks.

In the same way we use Chatbots in business environments or as customer facing but, unlike the previous ones ... they do not speak.

We have developed a solution to incorporate the voice to Oracle Digital Assistant in native Apps for Android and iOS, and that allows us to listen to the messages we receive from the bot and talk to it as if we were doing it with a person in real life.

To achieve that we need some prerrequisites:

• ODA (Oracle Digital Assistant) Platform.

• Native SDK for each platform (Android, iOS).

• Access to the APIs Text to Speech and Speech to Text.

First of all is we have to implement the Chatbot in ODA with the different features and use cases that you want to cover. In this article we are not going to deepen on this topic, so we start from the fact of having developed the Chatbot, and continue with the next part.

To begin with, we follow Oracle documentation to integrate native SDKs of the platform.

Next, we work with the classes provided to access the different events that occur in the conversation.

Once the events are configured, we will use the native APIs of the platform (Android, iOS) to use the voice synthesizer engine, so that we will hear the voice of Google Assistant and Siri respectively. In this way, the user keeps the experience acquired with his device, interacting with him as he has done so far in other scenarios, making it more pleasant.

To dictate to the Chatbot we can use the microphone symbol that the keyboard of the device provides or we can wait for the end of the speech to talk to you. The choice of one or the other will depend on the experience you want to give to the bot.

Finally, our recommendation is to generate different flows in the Chatbot, so that users can configure the behavior of the voice and thus have a user experience according to their preferences.

In the following videos you can see how this feature has been integrated for one of the bots we have done, in this case MatchBot that has been developed for the tenis tournament Mutua Madrid Open.