G-Particle

How Does a ChatBot Work?

/in /by

There are four main stages in a Chatbot:

How Does a Chatbot work?
how a chatbot works
  1. Interpretation: In this stage, the application analyzes the user’s input to determine the intention of the dialog. Once the intent is recognized, the Bot discovers all the entities from the dialog. For instance, in the above example, the purpose is identified as the greeting. There are no entities identified in this particular interaction.
  2. Dialog: Based on the purpose identified in the previous stage, the application decides on the type of response for the user. In the above picture, the Bot knows that the intent was to greet. Thus, the response drafted is, ‘Hello’  with the username variable.
  3. Data Retrieval: If the application requires any additional data to draft the final response, it either asks the user for further inputs or retrieves it from the relevant connected database. Once all the missing information is available, the Bot becomes ready to prepare the final reply for the user. As an exemplar, the above system fetches James.
  4. Response: Once all the necessary information is collected, the application prepares the final response and sends to the user.
pixabay

How to Build a ChatBot?

/in /by

Bots are a lot like modern web applications, living on the internet and use APIs to send and receive messages. Building a bot is a seven-step process, involving planning, design, development, testing, Go-Live, connection, and continuous assessment.

Chatbots – for human-like conversations with your clients across multiple platforms

All the seven stages are briefly described below:

  1. Planning: Before the actual coding of the bot, it is important to identify the needs the bot is being designed for, the goals it will achieve and the processes involved in accomplishing the goals.
  2. Design: The application is designed as per the plan prepared, covering all the use-cases.
  3. Development: Based on the design, it is important to identify the right development framework, bot services, environment, tools, and services, to implement bot web application.  
  4. Testing: A complex application, such as bot, requires rigorous testing to ensure that it is responding in the expected manner. Selection of right testing, debugging, interrogation, and web chat tools is a key.
  5. Go-Live: Once the application is ready, it is then published on the preferred web- service or data center.
  6. Connection: Bot is then connected to the messaging platforms, such as Facebook, Messenger, Kik, Skype, Slack, Microsoft Teams, Telegram, text/SMS, Twilio, Cortana, and Skype.
  7. Assessment: Based on the data gathered through various platforms, analytics are performed to identify opportunities and further fine tuning.

G-Particle is committed to utilize the best of the knowledge and technology to make the process of chatbot adotption easier for its customers.

Case Study: News Gets Personal with CNN developed Chatbot

/in /by

An innovative idea of using a chatbot to be closer and more personalized serving the market demand

Chatbots will ultimately have a profound impact on our digital lives. The technology enables both the intimacy of a one-to-one conversation as well as a mechanism for broadcasting a critical message at scale. That is precisely why CNN is being aggressive in this field; we believe these platforms can be a powerful way to deliver real-time, personal news to an audience. (said by Alex Wellen)

CNN is using bots to interact with users in a natural and conversational way. Subscribers to CNN on Messenger will receive a daily digest of top stories right within the Messenger app.  The bot will also recommend personalized content based on a user’s preferences and learned interests.  The experience gets more personalized with each interaction on Messenger. (full article)

Using chatbot is all about how to provide a more personalized experience to the users. Using a natural and conversational way of interaction with the users gives a better chance to organizations like CNN to achieve this matter.

Here is an article about the interview with Alex Wellen, CNN’s senior vice president and chief product officer after six months of chatbot experimentation.

From technology utilization standpoint of view, the main components of these solution can be summarized as below

  • Chatbot engine, responsible for receiving and sending messages to the users through various messaging platforms
  • NLP engine understands the user intent
  • Machine learning engine helps the whole solution to provide more specific and personalized responses with each interaction
  • Search engine, CNN in house content management system to provide the required content based on the user request and recommendations provided by the machine-learning engine
  • BIG Data and analytics responsible for analyzing, the user behavioral patterns and their likings based on various criteria, this is a unique advantage to CNN to plan their future strategies and investment plan closer to their market demand

SmartCity road-map of Yerevan city – Armenia

/in /by

A joint project of G Particle under supervision of UNDP

We are grateful to announce that G-Particle has been selected to participate in the preparation of the SmartCity roadmap document of Yerevan City and completed its assignment successfully 

Based on the strategical vision and priorities defined by leaders of the city and development program supervised by the United Nations, we could successfully identify the top priority projects which can satisfy the short-term and midterm needs of the citizens into three main categories. Read more

  • services for citizens and businesses 
  • efficient local government
  • infrastructure