What is a chatbot?Simulate human conversation for service

Chatbot definition

A chatbot is a software application designed to simulate human conversations with users through text or voice. Chatbots are also called virtual agents, interactive agents, digital assistants, or conversational AI, and are usually integrated into applications, websites, or messaging platforms to provide support to users without the use of real agents.

Liz Miller, vice president and principal analyst of Constellation Research, said: “Chatbot is a natural language processing application that allows users to conduct more and more voice-based conversations between users and digital interfaces, two-way conversations. .”

Chatbots originated from the user’s option menus, decision trees, or keyword-driven tools that look for specific phrases, such as “cancel my account.” The current iteration uses artificial intelligence and machine learning to create a more human-like experience.

With the improvement of artificial intelligence systems, the use of chatbots may increase substantially. According to data from Mordor Intelligence, from 2021 to 2028, the global chatbot market is expected to expand at a compound annual growth rate of 35%, reaching US$102 billion by then.

Chatbot example

Chatbots are used for customer and employee-oriented platforms and communication channels, such as websites, social media platforms or corporate communication systems, and are increasingly built into major enterprise software systems, such as customer relationship management, customer experience, and human resources And help desk platform.

They are used to answer common questions, and the natural language processing engine enables them to understand questions asked in unusual language. Chatbots can also be used to guide customers or employees to complete common tasks, or teach them how to use products and services.

Examples of chatbots include:

• The World Health Organization’s WhatsApp chatbot Health Alert can answer COVID-19-related questions in English, Arabic, Hindi, Italian, Spanish and other languages
• Bank of America’s Erica, a more advanced chatbot, is incorporated into the bank’s mobile application and can answer a variety of questions in English
• Eno, a chatbot similar to Erica, available on CapitalOne’s website and mobile app, and via email and text messages
• Domino’s Pizza’s Dom, helping customers process pizza orders

Top chatbot use cases

Su Jones, director of experience design at Nerdery, said that the most common use of chatbots is customer service. But there are also useful examples in healthcare, financial services, and several other industries.

Chatbots have also become a relatively popular tool in enterprises. For example, in HR, chatbots can help employees register for benefits or apply for vacation. IT chatbots can handle password reset requests or help diagnose connection issues. Chatbots can also be used in sales to suggest the best potential customers to call next, or in finance to answer queries about company performance data.

Bill Donlan, executive vice president of digital customer experience at Capgemini, said that when you start using chatbots, it is best to use them for common and relatively simple problems.

For example, when used for customer service, chatbots can handle some simpler requests, Donlan said, “allowing service agents to free up time to solve more complex problems.” He added, using it under appropriate circumstances and combining With manual support, chatbots can improve customer satisfaction, reduce costs, improve service availability, and improve overall efficiency.

But just because a company owns a chatbot does not mean that customers will use it.

“Most websites have it, but it is often overlooked,” said Evelyn McMullen, human capital and enterprise content management research manager at Nucleus Research. “In my opinion, they are not necessarily a substitute for human interaction, because the answer will be canned.”

Chatbot software

Major cloud vendors have chatbot APIs for companies to use when writing their own tools. There are also open source software packages available, as well as chatbots built into the main customer relationship management and customer service platforms.

Many companies also provide independent chatbots. According to Grand View Research, major chatbot vendors include 7.ai, Acuvate, Aivo, Artificial Solutions, Botsify, Creative Virtual, eGain, IBM, Inbenta, Next IT and Nuance.

“In the organizations we talk to, there is usually at least one chatbot platform that exists, even if they don’t necessarily know it exists,” said William McKeon-White, an analyst at Forrester Research. “In fact, the number of chatbots has exploded. The organizations we talk to sometimes have as many as 13 platforms, all of which compete internally.”

For example, developers may use open source chatbots to help them automate their work. McKeon-White said that they may have one in Microsoft Teams or Slack, or be integrated into other platforms, such as Jira.

Other business users may start using integrated chatbot features in platforms such as Salesforce or ServiceNow.

He said this means that all these platforms are competing for funding – and the attention of developers. “Organizations need to choose one, pool resources, and create a more effective platform for their users.”

Chatbots and artificial intelligence

Chatbots initially provided users with a simple selection menu, which then evolved to respond to specific keywords. “But humans are very creative in using language,” said McKeon-White of Forrester.

People who want to reset their password may say that they have forgotten the access code or have trouble accessing their account. “There are many different ways of expressing the same thing,” he said.

This is where artificial intelligence comes in. Natural language processing is a subset of machine learning, which enables the system to understand the meaning of written and even spoken language, even with many variations in wording. In order to succeed, chatbots that rely on artificial intelligence or machine learning first need to be trained using data sets. Generally speaking, the larger the training data set and the narrower the field, the more accurate and useful the chatbot will be.

“They are increasingly able to recognize the similarities of different discourses,” he said. “But we can all point out some bad examples we have encountered with chatbots because the conversation is difficult. It’s like taking a toddler to throw a dictionary at them and saying,’Go deal with this complex The problem.'”

He said that the greatest success lies in the company’s attention to the shortcomings of chatbots and continuous efforts to improve.