What is chatbot?

Over the phone, it is easy to tell if you are talking to a robot feeding you automated responses, but on the internet, it is much harder. That’s because of chatbots.

Chatbots are artificial intelligence (AI) software that can simulate a conversation. These robots’ primary purpose is to communicate with humans via text, voice, and touch. If you’ve used a customer support live-chat service on-line, it’s likely you were responding to a chatbot. Through advances in AI and Natural Language Processing (NLP), a subset of A.I., on-line chatbots effectively mimic human conversation. Chatbots are being adopted to automate processes like sales, marketing, lead generation, and customer service. A 2011 survey by Gartner predicted that 85% of our engagement with businesses would be done without interacting with another human being by 2020, and we’re getting close (Gartner Survey). There was a time when businesses hired a room full of people to provide customer service. Today, many of them rely on Facebook messenger bot, WhatsApp bots, and WeChat to reach out to their customers.

How Do Chatbots Work?

There are two main types of Chatbots:

Rules-Based Chatbots—The simplest type of Chatbot, it can only hold a conversation if the user says the right thing. People interact with these bots by clicking on buttons and using pre-defined options. To give relevant answers, these Chatbots require people to make a few selections. As a result, these bots have longer user journey, and they are the slowest to guide the customer to their goal. This bot is only as smart as it is programmed to be.

Smart Chatbots—Also known as AI Chatbots, they use Machine Learning, A.I, and Natural Language Processing NLP to understand the user. These Chatbots understand free language, but also have a predefined flow to make sure they solve the user’s problem. They can remember the context of the conversation and the user’s preferences. Over time, and by observing correct and incorrect answers, the machine gets better at understanding what the “right” answer is (think Apple’s Siri, Google Assistant, Amazon’s Alexa).

Fun fact: The famous Bank of America Chatbot, “Erica,” which launched in 2018, has served more than 10 million users since then and was able to understand close to 500,000 question variations by mid-2019.


Chatbot Challenges

With a projected worldwide market size of more than $1.3 billion by 2024, Chatbots will be a driving force for business communications. (Intellecty, 2019).

There are four primary challenges that contribute to the skepticism about the usefulness of Chatbots:

  • Chatbots’ inability to execute technical commands.
  • Chatbots are unable to process a customer’s intent, leading to misinterpreted requests and responses.
  • Chatbots lack conversational intelligence—that is, they often do not process implied nuances of dialogue, which results in inadequate conversation.
  • Chatbots are also unable to understand different accents or cultural meanings to process an accurate response.

Chatbots are an inciting target for hackers. As consumers get more used to interacting with Chatbots, there will be new opportunities for phishing, hacking, and general harm. Chatbot attacks range from man-in-the-middle attacks to polluting the communication channel (Microsoft Tay Chatbot Attacks).

Despite their limitations, more and more companies are investing in Chatbot technology because it will revolutionize the world. Chatbots are now mainstream in a diverse array of consumer applications and internal business functions. Consumers look forward to using Chatbots to purchase goods and services.

About Victor Allen

Our blog writer, Victor Allen, is a data scientist with over 20 years of subject matter expertise and robust knowledge in information technology, data science, machine learning, big data analytics, cyber security incident response and cyber security intelligence training. He provides advanced capabilities in Data Science methodologies and techniques of Data Extraction, Data Mining, Data Wrangling, Feature Extraction, Statistical Modeling, Predictive Modeling, and Data Visualization. Before joining Collabraspace in 2021, Mr. Allen worked as a data science technical lead developing and teaching foundational data science curriculum for the National Cryptologic Institute.