The history and use of conversational AI, including the ways conversational artificial intelligence evolves beyond the typical chatbot.
Editor’s Note: The article has been updated on Jan. 5, 2023 to include new data and information.
When people think of conversational artificial intelligence (AI) their first thought is often the chatbots they might find on enterprise websites. Those mini windows that pop up and ask if you need help from a digital assistant.
While that is one version, many other examples can illustrate the functionality and capabilities of conversational artificial intelligence technology.
What is conversational artificial intelligence, exactly, and how did it come to be? What does a conversation with artificial intelligence look like? And what impact will this technology have on business-consumer relationships?
The History of Conversational AI: From Chatbot to Present
The standard conversational AI definition is a combination of technologies — machine learning and natural language processing — that allows people to have human-like interactions with computers. To break it down further, let’s look at the evolution of conversational AI.
The Rise of the Chatbot
Chatbots made their debut in 1966 when a computer scientist at MIT, Joseph Weizenbaum, created Eliza, a chatbot based on a limited, predetermined flow. Eliza could simulate a psychotherapist’s conversation through the use of a script, pattern matching and substitution methodology.
Although Eliza could pass a restricted version of Turing test — a test that determines if a machine can display intelligent behavior indistinguishable from a human being — and fool people into thinking they were talking to another human, it was simply following rules and simulating the conversation with no real level of understanding.
New Natural Language Understanding
A decade later, Kenneth Mark Colby at the Stanford Artificial Intelligence Laboratory created a new natural language processing program called PARRY. Although it was the first AI program to pass a full Turing test, it was still a rule-based, scripted program.
Optimized Natural Language Generation
In 1995, Richard Wallace created the Artificial Linguistic Internet Computer Entity (ALICE). It used what was called the Artificial Intelligence Markup Language (AIML), which itself was a derivative of Extensible Markup Language (XML).
Like its predecessors, ALICE still relied upon rule matching input patterns to respond to human queries, and as such, none of them were using true conversational AI.
The Advancement of Conversational AI
What is conversational AI? It relies on natural language processing (NLP), automatic speech recognition (ASR), advanced dialog management and machine learning (ML), and can have what can be viewed as actual conversations.
Conversational AI also uses deep learning to continuously learn and improve from each conversation. It is flexible and able to jump from one topic to another, much like actual human speech and unlike traditional chatbots, which are limited to pre-defined scripts and rules and cannot respond with anything not originally inserted into its conversational flow.
“Rule based or scripted chatbots are best suited for providing an interaction based solely on the most frequently asked questions. An ‘FAQ’ approach can only support very specific keywords being used,” said Eric Carrasquilla, senior vice president and general manager of Digital Engagement Solutions at CSG.
“Conversational AI is ingesting the customer feedback and learning in real-time that value, which can be applied to the same question at a different point of a client’s journey.”
By using conversational AI chatbots, basic contact queries such as delivery dates, tracking numbers and shipping fees can be easily and quickly taken care of, while more complex or serious customer service inquiries can be passed on to live customer service representatives.
“The appropriate nature of timing can contribute to a higher success rate of solving customer problems on the first pass, instead of frustrating them with automated responses,” said Carrasquilla.
Related Article: How AI Is Shaping the Future of Customer Interactions
People Trust Conversational AI Solutions
What is an example of conversational AI? One top use today is to provide functionality to chatbots, allowing them to mimic human conversations and improve the customer experience.
An MIT Technology Review survey of 1,004 business leaders revealed that customer service chatbots are the leading application of AI used today. Nearly three-quarters of those polled said by 2022, chatbots will remain the leading use of AI, followed by sales and marketing.
Not surprisingly, a report from Capgemini, AI and the Ethical Conundrum, indicated 54% of customers have daily AI-enabled interactions with businesses, including chatbots, digital assistants, facial recognition and biometric scanners.
Nearly 50% of those customers found their interactions with AI to be trustworthy, up from only 30% in 2018. What used to be irregular or unique is beginning to be the norm, and the use of AI is gaining acceptance in many industries and applications.
It’s not just customers that are beginning to trust conversational AI. Those established in their careers also use and trust conversational AI tools among their workplace resources. Oracle and Future Workplace’s annual AI at Work report indicated that 64% of employees would trust an AI chatbot more than their manager — 50% have used an AI chatbot instead of going to their manager for advice.
Twenty-six percent of those polled said bots are better at providing unbiased information and 34% said they were better at maintaining work schedules. Not only that, but 65% of employees said they are optimistic, excited and grateful about having AI bot “co-workers” and nearly 25% indicated they have a gratifying relationship with AI at their workplace.
The Washington Post reported on the trend of people turning to conversational AI products or services, such as Replika and Microsoft’s Xiaoice, for emotional fulfillment and even romance.
With the isolation, separation and loneliness the pandemic brought with it, and the advances in artificial intelligence, many people have found that AI-based chatbots, and even AI voice bots, offer a user experience that fulfills the need for communication with other humans. In fact, Xiaoice has 10 million active users in China.
Conversational AI Is Trusted — but Is It Safe?
People trust conversational AI solutions and they find the technology helpful when they need to search for information. But does that mean its safe to use.
With the two examples of conversational AI above, where people have private conversations with a bot, perhaps even share personal information, the question of privacy and security might come to mind. How safe is conversational AI to use?
Like most things, conversational AI is as safe as it’s built to be. Users not only have to trust the technology they’re using but also the company that created and promoted that technology. Finding out if a specific conversational AI application is safe to use will require a little bit of research into how the bot was made and how it functions.
You’ll want to look for three things when it comes to finding a safe AI bot:
- End-to-end encryption: Encryption increases application security by ensuring no one else but the sender and receiver can access a conversation. End-to-end encryption is also necessary to comply with the General Data Protection Regulation (GDPR) in the European Union.
- Authentication processes: These are processes that help verify a person’s identity — like using your thumbprint to gain access or sending a verification code to your phone that you then type in.
- Privacy policies: Companies will have written privacy agreements on their website or application that covers how they collect information and what they do with it. The ideal application would not share or sell private information to any other entity.
Conversational AI users should also ensure they have a fundamental understanding of internet safety measures, including:
- Using strong passwords and changing them regularly
- Keeping applications up-to-date
- Using websites that start with “HTTPS” rather than “HTTP”
- Using an ad-blocking extension to block pop-ups and spam
- Using a trusted antivirus that’s kept up-to-date
- Enabling multi-factor authentication for important accounts
- Avoiding sketchy downloads that could lead to a virus
Related Article: How Will Conversational AI Transform Customer Experience?
Businesses (and People) Rely on Omnichannel Conversational AI
Traditional chatbots are text-based. They’re typically found on only one of a brand’s channels — usually a website. They aid in customer service conversations and can improve the overall customer experience.
Conversational AI solutions, however, are omnichannel. They can be accessed and used through many different platforms and mediums, including text, voice and video.
“The pairing of intelligent conversational journeys with a fine-tuned AI application allows for smarter, smoother choices for customers when they reach out to connect with companies,” Carrasquilla suggested.
Among other common conversational AI examples is the digital assistant — think Cortana, Google Home, Amazon Alexa and Siri.
According to a report from National Public Media, 24% of people over 18 (around 60 million people) own at least one smart speaker, and there are around 157 million smart speakers in US households. Smart speakers account for approximately 69% of voice assistant users.
The use of smart speakers and virtual assistants has facilitated the acceptance of conversational AI in the household. According to Google, 53% of people who own a smart speaker said it feels natural speaking to it, and many reported it feels like talking to a friend. Several respondents told Google they are even saying “please” and “thank you” to these devices.
Some call centers also use digital assistant technology in a professional setting, taking the place of call center agents. These digital assistants can search for information and resolve customer queries quickly, allowing human employees to focus on more complex tasks.
Chris Radanovic, a conversational AI expert at LivePerson, told CMSWire that in his experience, using conversational AI applications, customers can connect with brands in the channels they use the most.
“Intelligent virtual concierges and bots instantly greet them, answer their questions and carry out transactions, and if needed connect them to agents with all of the contextual data they’ve collected over the course of the conversation,” he said.
Related Article: Can Conversational AI Improve the Online Retail Experience?
Conversational AI Facilitates Hyper-Personalization
“Hyper-personalization combines AI and real-time data to deliver content that is specifically relevant to a customer,” said Radanovic. And that hyper-personalization using customer data is something people expect today.
Radanovic emphasized that consumers and brands are embracing conversational AI because it provides personalized experiences that are also much quicker and more convenient than traditional ways of interacting with businesses. Customers do not want to be waiting on hold for a phone call or clicking through tons of pages to find the right info.
According to Radanovic, conversational AI can be an effective way of eliminating pain points in the customer journey.
“A giant source of frustration for consumers is repeating information they’ve already shared, like re-confirming a phone number or having to re-explain a problem to multiple agents.
As brands adopt tools that allow conversational AI to connect customer data, said Radanovic — like connecting conversation histories with previously stated intentions — the conversations they have with customers will feel more personalized.
Conversational AI Is Part of Our Daily Lives
Traditional chatbots still have their uses. They’re great at answering common questions with a standard script. But when you need a little more complexity, you need conversational AI.
Conversational AI models, powered by natural language understanding and machine learning, are not only very effective at emulating human conversations but they have also become a trusted form of communication. Businesses rely on conversational AI to stimulate customer interactions across multiple channels. The tech learns from those interactions, becoming smarter and offering up insights on customers, leading to deeper business-customer relationships.
But these bots go beyond pure business use. People use these bots to find information, simply their routines and automate routine tasks. Ultimately, they’ve become an extension of people’s daily lives.
