<img alt="" src="https://secure.leadforensics.com/150446.png " style="display:none;">

Meet Zesty, our very own Virtual Assistant

Chelsy Bhandari Jun 10, 2019

Machine Learning Natural Language Processing customer service AR/VR

I am reminded of Tony Stark’s dialogue from the movie 'Avengers' about his digital assistant JARVIS. "Started out as just a natural language UI. Now he (referring to JARVIS) runs the Iron Legion. He runs more of the business than anyone besides Pepper." - Tony Stark to Bruce Banner.

Right from a sci-fi technology from your favorite movie, to a working model in today's industries, robots have come a long way and are no more a thing of fiction. Robots are increasingly being utilized in almost every industry. For instance, robotic arms and rovers are being used in the Space industry, welding and painting robots are a common sight in the manufacturing industry and logistics sector has started using robots for delivery.

Robotics – A convergence of the physical and digital has changed the way we work while adding never-before-seen value to performance and productivity over the last 15 years.

Robotics automates different functions across the social and enterprise spectrum, reducing the costs of goods and freeing up employees and citizens to become more efficient. The impact of this robotic revolution is growing and is now being felt at multiple channels.

There has been constant research to prove and establish the human-like peculiarities in an Artificial Intelligent (AI) agent which in simple terms may be called as a robot - A humanoid with Artificial General Intelligence (AGI) – one of the top achievements in the human history of innovations.

We at e-Zest, try continuously to be at the forefront of innovations using emerging technologies to provide state of the art solutions. Here’s a brief story of how and why we are developing Zesty - Our very own artificial intelligent robot.

Why Zesty?

By developing a natural language UI, we aspire to open a plethora of possibilities. Robots can reduce human labor and error, can improve work quality and make any process redundant over time to achieve the maximum tolerance of free services.

“Where there is data, there is data science”

Conversational experience matters most in human social life. Technology today has evolved at a stage where tasks of a receptionist like greeting, managing logs, and answering Frequently Asked Questions (FAQ’s) are now possible fluently with minimal errors handled by robots.

A greeting is an act which initiates the communication and the type of greeting can suggest a type of relationship both parties have. In almost every organization, there is a receptionist who greets people who manage the data of visitors. We are trying to automate these tasks through Zesty.

How Zesty will work

A simple daily task of talking can be arduous to a machine. But today, we can build conversational agents who can talk and respond like humans. Live examples are 'Google Assistant' and 'Siri' on your smartphones.

On similar lines of a smart assistant, Zesty will recognize visitors and employees with face recognition models which are based on the convolutional neural networks to give a personalized experience. We will have control over Zesty with our business logic to perform tasks like meeting scheduling, information gathering, etc.

Zesty will be able to take your information verbally in an interactive conversation similar to a human. This will be achieved with user utterances and intent matching. Also, in addition to this, we can attach the identity card printing mechanism to print new IDs for any events spot registration.

Here are some of the use cases as follows:

  • Use case 1:
  • If a person is coming in front of Zesty for the first time then the face recognition model will identify that person as unknown. And Zesty will ask the reason for visit or any custom questions on which it has been trained.
  • Conversation:
  • Zesty - Thank you for giving your information, your ID will be provided you shortly.
  • Unknown user - Person will give his information to Zesty. (And if by chance Zesty misses an entity then it can automatically go into fallback intent which asks again the missing entity)
  • Zesty - Please tell me your name, birth date, experience, etc.
  • Unknown user - Okay!
  • Zesty - Yes, please answer some simple questions.
  • Unknown user - can I have new identity card please?
  • Zesty - Yes sure! How may I help you?
  • Unknown user - Hello I am new here, can you help me?
  • Zesty - Hello there! Who are you?
  • Use case 2:
  • If a person is coming in front of Zesty is known or in other words, Zesty has its facial data beforehand then it can recognize its face and provide custom greeting and services.
  • Zesty - Hello (name of person) nice to see you, do you want my help.
  • Known person - No thank you I am good.

Behind the scene - For the techies

NLP capabilities are achieved by Natural language Engine (NLE). Face recognition is a continuously evolving field where research is going on rigorously. Researchers are constantly trying to mimic and achieve human-like accuracy in recognizing faces.

Point to be noted is, Face Detection and Face Recognition are two different things. We have achieved and, in some cases, have surpassed the humans in detecting the faces but the scope of improvement is in face recognition. Currently, we are using a custom convolutional neural network on which we train faces. The training will be totally parallel.

We have used Python as a base language but our server which will handle business logic is a Node Server. This server can be used to add additional functionalities like weather, traffic updates, and any database integrations.

And Zesty can be a virtual robot personality on screen or a physical Robot. A virtual robot will be a 3D model which will be rendered in sophisticated software. This character is customizable to suit any organizational needs. Along with this, a dashboard displaying text conversation and animations can be added for a more immersive experience.

And a physical robot can have a 3D printed structure or a lightweight carbon fiber structure. Onboard microprocessors will be used. which will have onboard parallel processing capabilities which will boost the performance of machine learning algorithms and workflows.

Zesty and others

A virtual agent which is being built for enterprise can be used in home automation and for security purposes.

Yes, today there are a lot of home automation solutions out there, but we are thinking to introduce a personality to your automation and security services.

A near to human level personality can cheer the mood of the users. The more natural social interactions between humans and robots, the more meaningful the engagement. This is precisely what Zesty will be – a friendly companion in your house or organization, with personalized skills.

e-Zest is a leading digital innovation partner for enterprises and technology companies that utilizes emerging technologies for creating engaging customers experiences. Being a customer-focused and technology-driven company, it always helps clients in crafting holistic business value for their software development efforts. It offers software development and consulting services for cloud computing, enterprise mobility, big data and analytics, user experience and digital commerce.