Checking out chatbots: simulating human conversation for service

A chatbot is a software application designed to simulate human conversation with users via text or speech. Also referred to as virtual agents, interactive agents, digital assistants, or conversational AI, chatbots are often integrated into applications, websites, or messaging platforms to provide support to users without the use of live human agents.

“Chatbots are an application of natural language processing that allows typically text- but increasingly voice-based conversations, bidirectional, between a user and the digital interface,” says Liz Miller, vice president and principal analyst at Constellation Research.

Chatbots originated as menus of options for users, decisions trees, or keyword-driven tools that looked for particular phrases, such as “cancel my account.” Current iterations use AI and machine learning to create a more human-like experience.

As AI systems improve, chatbots are likely to see massive growth in use. According to Mordor Intelligence, the global chatbot market is expected to expand at a compound annual growth rate of 35 per cent from 2021 to 2028, when it will reach US$102 billion.

Chatbot examples

Chatbots are used on customer- and employee-facing platforms and communication channels, such as websites, social media platforms, or enterprise communication systems, and are increasingly built into major enterprise software systems such as customer relationship management, customer experience, HR, and help desk platforms.

They are used to answer common questions, with natural language processing engines enabling them to understand questions posed with unusual wordings. Chatbots can also be used to guide customers or employees through common tasks, or teach them how to use products and services.

Examples of chatbots include:

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

Top chatbot use cases

The most common use of chatbots is in customer service, says Su Jones, director of experience design at Nerdery. But there are also use cases in healthcare, financial services, and several other industries.

Chatbots have become relative popular tools within the enterprise as well. In HR, for example, a chatbot can help an employee sign up for benefits or request time off. An IT chatbot can process a password reset request or help diagnose a connectivity issue. Chatbots can also be used in sales to suggest the best prospects to call next, or in finance to answer queries about corporate performance numbers.

When starting out with chatbots, it’s best to target common and relatively simple issues for their use, says Bill Donlan, executive vice president for digital customer experience at Capgemini.

For example, when used in customer service, chatbots can take on some of the easier requests, Donlan says, “freeing service agents for more complicated problem solving.” Used in proper situations, and combined with human support, chatbots can result in higher customer satisfaction, reduced costs, greater service availability, and better efficiency overall, he adds.

But just because a company has a chatbot doesn’t mean customers will use it.

“Most websites have it, but it’s often ignored,” says Evelyn McMullen, research manager for human capital and enterprise content management at Nucleus Research. “In my opinion, they’re not necessarily a replacement for human interaction because the answers are going to be canned.”

Chatbot software

The major cloud vendors all have chatbot APIs for companies to hook into when they write their own tools. There are also open source packages available, as well as chatbots that are built right into major customer relationship management and customer service platforms.

Standalone chatbots are also available from a number of companies. According to Grand View Research, key chatbot vendors include 7.ai, Acuvate, Aivo, Artificial Solutions, Botsify, Creative Virtual, eGain, IBM, Inbenta, Next IT, and Nuance.

“In the organisations we talk to, there’s usually at least one chatbot platform present, even if they don’t necessarily know it exists,” says William McKeon-White, analyst at Forrester Research. “In fact, there’s been a proliferation of chatbots. Organisations we talk to sometimes have up to 13 platforms all competing internally.”

For example, a developer might adopt an open source chatbot to help them automate their work. They might have one in Microsoft Teams or in Slack, or integrate into other platforms, such as Jira, says McKeon-White.

Other business users might start using the integrated chatbot capabilities in platforms such as Salesforce or ServiceNow.

That means that all these platforms are competing for funding — and for attention from developers, he says. “Organisations need to choose one, to centralise resources for the creation of more effective platforms for its users.”

Chatbots and AI

Chatbots originally started out by offering users simple menus of choices, and then evolved to react to particular keywords. “But humans are very inventive in their use of language,” says Forrester’s McKeon-White.

Someone looking for a password reset might say they’ve forgotten their access code, or are having problems getting into their account. “There are a lot of different ways to say the same thing,” he says.

This is where AI comes in. Natural language processing is a subset of machine learning that enables a system to understand the meaning of written or even spoken language, even where there is a lot of variation in the phrasing. To succeed, a chatbot that relies on AI or machine learning needs first to be trained using a data set. In general, the bigger the training data set, and the narrower the domain, the more accurate and helpful a chatbot will be.

“They’re increasingly able to identify the similarity of different utterances,” he says. “But we can all point to some bad instances we’ve had with chatbots, because conversations are hard. It’s like taking a toddler and throwing a dictionary at them and saying, ‘Go handle this complex problem.’”

The biggest successes, he says, are when companies pay attention to where the chatbot is missing the mark and continually work to improve it.