|
Listen : Audio version of this article
|
Conversational agents or chatbots that were initially thought to primarily be a job creator have now become a way to declutter customer requirements and queries with the help of machines. Several industries have integrated this form of machine learning in their systems to address customer requirements efficiently.
According to a report on the Chatbot market by Reports and Data, the global chatbot industry is expected to grow to a valuation of USD 10.08 Billion by the year 2026. Chatbot technology is predominantly run by natural language processing (NLP), the same technology that forms the foundation of a virtual assistant like the ones used by Google, Apple for its Siri, or by Microsoft for its Cortana.
The technology processes the text received by it by what is known as ‘parsing’ and responds with the help of a complex series of algorithms that enable it to interpret and identify their requirements, and determine the most appropriate response according to the information received. Some advanced chatbots have a very authentic conversational experience, which can create a dilemma for the listener about the speaker being a machine or a human being.
The development of the first chatbot is credited to Professor Joseph Weizenbaum of the Massachusetts Institute of Technology in the year 1966. He designed a program called ELIZA, which could simulate a psychotherapist, with the help of keywords and pattern matching to converse with the user.
Chatbots have been called by many names, including talk bot, chat robot, chatterbot or chatterbox, and academically referred to as Artificial Conversational Entity (ACE). Even if they can speak like humans, people can be wary of speaking to a machine, which is why developers prefer to give human names to chatbots, for instance, Siri and Alexa, or even give them human-like avatars that pop up when interacting with a user.
We are continually trying to improve the technology in these chatbots to make the machine sound more human. A research performed by a team at Penn State concluded that chatbots that had human features but poor interactive capabilities were not well received by the users. However, they observed that people were more responsive to a less-interactive chatbot that had lesser anthropomorphism i.e., human-like features.
Co-author S. Shyam Sundar explains that high interactivity would be indicated by swift responses that correspond with the user’s queries and has a threaded exchange that is easy-to-follow. The study suggests that people are pleasantly surprised when a chatbot that has low anthropomorphism is more interactive, but when they converse with a chatbot with higher anthropomorphic aspects, it increases their expectations for interactivity, leaving them equally disappointed.
However, higher interactivity is more sought-after even if the chatbot is less human-sounding, which can be accomplished by minuscule changes in the dialogue or by acknowledging the previous statement by the user.
Modern chatbots are used when simple interactions are needed with a limited number of responses. This mostly encompasses customer service or applications in marketing, wherein the chatbots are equipped with information about products and services, and in some cases, even company policies. To maintain the smooth operation of the entire process, the user is typically provided with access to a human operator, in case the user’s queries are not entirely resolved by the chatbot.
Chatbots are mostly used for online services and in messaging apps, but they are also included in the many other operating systems as intelligent virtual assistants. Over time, chatbots evolved from responding textually to being able to fulfill the command of the user. For instance, Amazon’s Alexa or other smart home assistants, which are interwoven with the electronic system of the house and can perform numerous functions in response to specific keywords said by the user.
Even though some reports are questioning their efficacy in keeping customers satisfied, chatbots have established their place in several industries. Chatbots were predicted to be vital for some activities like customer service and marketing, but healthcare has surprisingly become one of the largest sectors where chatbots are prevalent.
Over the years, it has successfully incorporated machine learning for most of its processing activities to analyze data and maintain electronic records, but its capabilities have moved beyond clerical services. Currently, chatbots are not necessarily meant to fulfill customer requirements anymore; they are also designed to offer companionship in certain cases.
For instance, the chatbot named ‘Endurance’ has been created to provide companionship to dementia patients. People with dementia experience short-term memory loss and struggle to hold up conversations. This data is also examined to track the progress of memory loss to see the effects of therapy, among other things.
Similarly, the ‘Casper bot Insomnobot-3000’ has been developed to be a companion to a person who has insomnia when they do not have any person to talk to during the hours they are awake at night. Then, there is the chatbot named ‘MedWhat,’ which has been created to help accelerate the process of medical diagnosis.
The above examples related to the healthcare applications of chatbots show that even in the sector, where human interaction is almost mandatory, chatbots have found a way to perform more than just mechanical tasks. Its rising significance in the healthcare industry might be the biggest indication of the increasing anthropomorphism of the technology.
One thing is evident: chatbots are here to stay, whether we like it or not. Chatbots are now being tailored according to the requirements of different industries. For instance, ‘Roof.ai’ is a chatbot for marketers in real estate to automate their conversation with prospective assignments through social media, keeping up with the time constraints that are associated with sales in any industry. ‘SimbiBot,’ created by the University Tertiary Matriculation Examination (UTME), helps students prepare for exams by providing a database of the past questions.
Then, there is ‘Spot,’ a chatbot that allows users to anonymously report sexual harassment and discrimination at the workplace, which is dedicated to a more social cause. If we keep aside our fear that the machines will someday supplant humans, these automated bots are a boon for sorting through data and addressing the requirements of users in lesser time and with minimal human input. Companies, big and small, have incorporated the technology in their systems and given these machines access to information that allows them to replace human interaction that was otherwise considered necessary, although there is still a lot of improvement and advancement we can expect to see in the future.