Generative AI: commodity or core business?
04 March 2024
The choice between buying or developing generative AI in-house depends on the use, the data and the competitive advantage sought. Eleven proposes a methodology for…
Data science
A Voicebot is a conversational robot capable of engaging with a user through voice commands. In other words, a Voicebot is a virtual agent capable of understanding, interpreting, and analyzing requests expressed by a human to respond vocally in natural language—the way humans speak. A Voicebot is constructed using artificial intelligence, encompassing technologies that enable the robot to learn and understand human language and the intentions conveyed by the words used.
When we talk about intelligence, we often think of the brain. The “brain” of the robot is a decision tree—an instrument that, based on received information, allows us to select the appropriate response from multiple solutions. The robot’s intelligence is thus limited to understanding the received message and choosing the appropriate response based on its decision tree. It is a far cry from an AI like ChatGPT, with which we can engage in open conversations.
For example, well-known Voicebots include Siri for iPhone users and Alexa for Echo users.
To illustrate, let’s delve into an environment that touches every French citizen and where Voicebot technology is thriving: insurance. Over the past three years, especially since the Covid-19 pandemic, Voicebots have seen widespread deployment among insurers such as Axa, GMF, or Maaf, each adopting one or more Voicebots.
The deployment of Voicebots is driven by the increase in telephone calls during the Covid-19 period and the shortage of personnel in call centers. As an indication, a Voicebot deployed in a major insurance group can handle up to 400,000 calls per week.
From a client perspective, Voicebots offer an immediate response to their needs. A prompt response tends to satisfy the client, thereby enhancing the company’s image.
From an employee perspective, Voicebots save time, allowing them to focus on less repetitive tasks with potentially more added value. Employees are no longer stuck in repetitive tasks and can dedicate more time to more crucial actions.
The amalgamation of AI, robots, and work raises concerns about the ongoing debate between robots and humans. In scenarios where robots substitute for humans, the latter lose their jobs. However, the efficiency of the robot is heavily reliant on a human surveillance and improvement team; without it, the robot may be ineffective and could even deteriorate customer relations.
A Voicebot doesn’t speak human naturally. Like a child learning to speak, it needs regular exposure to new words to expand its vocabulary and comprehend what is said. However, the conversations it is exposed to are often of the same nature, leading to a decrease in the breadth of its vocabulary. At this point, human intervention becomes indispensable to train the Voicebot by teaching or reminding it of words.
Additionally, similar to a child, a Voicebot can have misunderstandings. Human intervention is necessary to correct its interpretation. Unlike a child who grows into an adult, a robot remains in a perpetual state of language learning, requiring continuous monitoring throughout its use.
Added to this is the necessary surveillance for any machine. Despite its ability to learn, the robot is not yet capable of repairing itself in case of bugs or even correcting its decision tree. Human action in its creation is therefore crucial.
The Voicebot doesn’t come to replace the human but acts as a tool. On one hand, it demands attention, which must be provided by a person. On the other hand, the robot can only fulfill standardized tasks within a predefined process. Hence, a collaborator remains essential to handle more complex or unusual tasks.
However, the Voicebot effectively executes an action that was previously performed by a less qualified employee in a call center. This employee often finds themselves replaced by a robot, but it’s precisely in this context that the concept of role evolution gains significance. Thanks to the Voicebot, this collaborator is freed up, allowing them to dedicate time to upskilling.
Through the use of a Voicebot, some insurance companies have increased the number of training sessions for these types of collaborators, enabling them to evolve more rapidly and perform actions that robots are not capable of. This strategy aims to enhance the qualified salary mass, especially in roles exposed to clients, ultimately improving service quality.
Moreover, it could be imagined that these same collaborators, having a knowledge of the field, could be trained in Voicebot supervision and transition to more technical positions overseeing Voicebot projects, roles often lacking in large structures.
The Voicebot eliminates positions to create new ones and allows time for training. It depends on the company’s strategy to leverage the Voicebot to improve its internal resources.
Share
Sur le même sujet