World Voice Day: The main way to interact with your mobile or computer will be through voice aids | Technology

World Voice Day: The main way to interact with your mobile or computer will be through voice aids | Technology

Smartphone manufacturers compete to integrate the algorithms of their devices into the routines of our day to day. Pixabay

For 24 years, the Israeli company VocalTec offered the first application to speak on the Internet (VoIP). The unstoppable irruption of intelligent speakers have relaunched the applications They use speech to perform all kinds of actions taking advantage of the network: from legal consultations to contracting trips or purchases. The field is as wide as the rest of the actions that are already carried out on a regular basis from the mobile or the computer, but which in the immediate future will be carried out talking to our devices. The global VoIP market, valued at 77,400 million dollars in 2018, is expected to reach 93.2 billion dollars by the end of 2024, with an annual growth rate of 3.1% between 2019 and 2024, according to Global Market data. Insights disseminated by the telecommunications operator Masvoz.

The speakers are intentionally less intelligent than they appear. They have emerged as an "operating system" and few applications, just as mobile telephony did, with the hope that it is the market that will feed your brain. No trading company of these platforms that responds to commands by voice could embark on the myriad of possibilities they offer. But the market is open to developers and the answers are already underway. Alexa, Siri, Hey Google, Bixby or Cortana are to get in touch with the robot that responds to the demands, but not to solve them on their own, except in a few cases.

"But anyone can make a voice application with very little investment," says Carlos Rodríguez, CEO of Chatbot Chocolate, a company present in Spain, Chile, Mexico and Colombia and which develops programs for users to access business services. just with asking. The cost ranges between 10,000 euros for basic developments up to 100,000 euros for more complex systems.

The first attempts to respond to voice commands on the Internet failed because they tried to imitate the complex system of human communication. Those attending stores and services were unable to finish a conversation and their results showed more failures than solutions. But Facebook changed the optics of attending to the conversations, Rodríguez recalls, incorporating automated responses that limited the dialogue to an environment understandable by the machine and useful enough for the user. The open interaction was limited to guided sequences that took into account the most common interactions.

The model was copied and incorporated into all the messaging platforms, which monopolize almost all the use of the mobile. Now, computer chats identify keywords and focus their solutions.

The main way to interact with the mobile or computer will be through voice aids

"As users, we spend 90% of our time in the smartphone in only three applications, with instant messaging being the main one. These factors, together with the fact that companies need to automate their customer service as much as possible, converge in the phenomenon and current boom of chatbots. Its mission is simple: be part of our day to day to resolve questions, questions or purchases, and help us get relevant information instantly without waiting or endless lines in phone calls. The recent emergence of smart loudspeakers in our homes also adds to the appeal of the voice when interacting with chatbots, "writes Julio Prada, CEO of Inbenta in Spain and EMEA.

The keys to this expert are two. The first and most difficult is the combination of artificial intelligence and the Natural Language Understanding. None of these developments, despite the efforts of companies such as those mentioned, Google Dialog flow, Amazon Alexa, Microsoft or IBM Watson, are sufficient for very complex environments, although the last of these companies has dared to present a robot that competes with a champion of debates.

There comes the second key: the botmaster, responsible for teaching, managing and maintaining the chatbot. But the probabilities of interaction are so wide that the task is titanic. A business needs an average of between 200 and 800 "intents" (intentions) to cover a use case that has to be multiplied by a minimum of 20 variations, explains Prada in his writing.

Carlos Rodríguez believes that, despite this difficulty, the technology of the chatbot is more than ready to "apply it to any situation that is resolved with a conversation". "80% of interactions do not need artificial intelligence," he warns. And the field is as big as any action that runs on a mobile or on a computer keyboard.

Companies have begun to offer these applications for the resolution of basic administrative procedures. Chatbots are already present in information offers, leisure, gastronomy or travel. Even in the assistance and surveillance system to elderly people, who can order that they call a relative if they have a domestic mishap without always having to carry the red remote assistance button.

Banks also apply them, but for informational purposes since the impossibility of guaranteeing privacy and security limits their use in transactions with money.

But the possibilities are huge. From receiving a private class at home to consulting with a robot lawyer or doctor basic questions. From these experiences there are already practical examples in China, Australia and California.

One of the spaces with the greatest implementation of voice commands is the car. According to a recent study in the United States, 67.5% of drivers routinely use them to make a call (73%), ask directions (50%), send a message (41%), or select the music or station of radio.


Source link