Evolution of next gen chatbots: How natural language processing is changing the game

With Artificial Intelligence (AI) revolutionizing all walks of business and social life, enterprises are plugging in conversational AI to enhance cross channel customer experience.

With Artificial Intelligence (AI) revolutionizing all walks of business and social life, enterprises are plugging in conversational AI to enhance cross channel customer experience. Gartner predicts that more than 50 percent of enterprises will opt for chatbots as the preferred channel over traditional mobile apps.

Representational image.

Representational image.

The technology at the core of the rise of the chatbot is natural language processing (NLP). NLP infused chatbots, designed to mimic human conversations, have gone through the peak of Gartner hype cycle expectations, the trough of disillusionment and are heading towards enlightenment and productivity.

With more people using chat to communicate, chatbots have gained a lot of importance and are fast becoming a trend for digitally empowered consumers. NLP provides a way for chatbots to understand and interpret the user in their own language to offer a richer conversational experience.

Basic building blocks of chatbots

Chatbots have a front-end conversational interface that connects to a variety of channels such as Facebook Messenger, Slack, Skype etc., with NLP to parse user messages and conversational logic.

Firstly, chatbots need to understand user input. It can achieve this through a basic technique of pattern matching or advanced intent classification technique that uses machine learning (ML).

A pattern matching technique needs a list of possible input patterns. They are easy to read and maintain. The problem is that they are built manually and do not scale in real use cases.

An intent classification approach relies on machine learning techniques and you need a set of user utterances to train a classifier.

Once it understands what the user says, it can generate a response, based on the current input and the context of the conversation. The simplest way is to have scripted static response for each user input. Another approach, would be to use a knowledge base to generate a dynamic response based on the context.

Chatbots come with different level of intelligence

The way we use chatbots and the method they use to respond to the user is changing. Initially, they would provide predetermined responses that are handcrafted as explained in the earlier section.

A more advanced approach is to use the deep learning technique to train a generative model and get a list of potential responses and then score them to choose the best suited response. Deep learning algorithms mean that chatbots are able to learn from every interaction, and incorporate that feedback so that their performance is continually improved. This cycle of continual improvement means that they are refining themselves over time.

Chatbots based on generative models are smarter but you need millions of examples of meaningful conversational data to attain a decent quality. This is an extremely promising area of current research which gives hope that the bot will get better at imitating humans.

Understanding the vendor landscape before you implement

In practice, chatbots are either goal oriented or conversational in nature. A goal-oriented or transactional chatbot helps a user perform specific tasks such as booking a ticket or ordering food.

A conversational chatbot, on the other hand, does not necessarily have a well-defined intention. It focuses on having an open domain conversation with the user and does not have to remember all the context of conversation.

To create a successful chatbot, the market is flooded with an array of platforms and tools, having different complexity levels, conversational intelligence and integration capabilities. Leading technology giants like IBM, Google, Microsoft, Amazon, Facebook are investing in this space to provide NLP, bot & AI frameworks, deployment platforms and as-a-service platforms.

The most common ones include Dialogflow (Google, ex API.ai), Amazon Alexa, Luis.ai (Microsoft), Wit.ai (Facebook), and IBM Watson.

There are a lot of aspects to consider when implementing a chatbot. It is important to find the platform that fits your particular need. Open source NLP libraries such as spaCy and AllenNLP are available for companies who prefer a customized chatbot solution.

In conclusion, chatbots are poised to revolutionize user interface design. As chatbots and intelligent automation are on the rise, enterprises should look at NLP infused chatbots to drive cost saving, operational efficiencies, and enhanced customer experiences throughout their businesses.

The author is the principal architect at Persistent Systems.

Loading...




Top Stories


also see

science