Editor’s Note: This article has been updated to ensure it has the most recent information and data.
Most of us have had online interactions with chatbots that were less than satisfactory, leaving us to resolve our issues another way. But today’s AI-based chatbots can have full-blown conversations that leave people feeling like they finished a conversation with a person.
Conversational AI that uses natural language processing, automatic speech recognition, advanced dialog management, deep learning and machine learning is likely to pass the Turing Test — a test that determines if a computer can think like a human — and provide a more realistic experience than traditional chatbots.
According to a report from MIT Technology Review, approximately 90% of those polled reported measurable improvements in their resolution speed for complaints. More than 80% reported enhanced call volume processing using AI, and the same number reported measurable improvements in service delivery, customer satisfaction and contact center performance.
“The Future of Work,” a report from Robert Half (registration required for report), revealed that 39% of IT leaders use AI or machine learning. An additional 33% indicated that they plan to use AI within the next three years.
Uses of Conversational AI
Most of our current conversational AI tools use weak or narrow AI and are limited to a specific field of tasks. Strong AI, as defined by the University of California, Berkeley, creates a human-like consciousness. It can solve many problems and perform tasks at an intellectual capability equivalent to a human’s.
Current conversational AI tools, though limited, are valuable for businesses of all sizes. Many provide faster customer response times, greater customer satisfaction and cost savings. They also have plenty of practical applications.
Online Customer Support
Conversational AI tools can reduce the number of people needed to ensure a positive customer journey. They respond to frequently asked questions for various industries, including banking, airlines and online businesses.
As they improve in efficiency, conversational AI tools change how businesses engage customers, particularly on the internet and through social media.
Conversational AI tools can perform many HR processes, including onboarding, employee training, answering employee questions and updating employee information.
AI tools can improve healthcare services to make them more accessible and affordable. They can enhance various administrative processes, including helping patients file claims and receive reimbursements faster.
Digital Personal Assistants
Many people have devices that use the Internet of Things (IoT), such as Alexa, Google Assistant or Siri. These conversational AI tools use customer input to improve responses to various questions and requests, such as product pricing or availability.
Some consumers may have internet-connected devices that allow them to respond to voice commands, such as refrigerators, ovens or lighting systems. The future, however, will see many IoT devices in homes.
Related Article: 4 Ways Conversational AI Is Improving the Customer Experience
Chatbots and AI Have Evolved
The chatbots we’ve all interacted with online are just one example of “old school” chatbots. Another is the interactive call routing voice systems many companies, such as doctor offices and utility providers, use to direct calls.
These systems provide basic and predictable information: the hours a business is open, an address or a website domain. A call system might list out departments and direct callers to select a number. Anything beyond that might be a challenge.
People have also created old-school chatbots for personal reasons. Wired magazine debuted a story in 2017 about a man who recorded conversations with his dying father for months. He used the recordings to create a conversational AI chatbot featuring his father’s voice. The idea was to develop a form of automated immortality.
Chatbots can also be predictive and personalized, with more complex, fluid responses similar to human decision-making. Aside from having access to a customer’s previous interactions through customer relationship management software, conversational AI can observe user-specific traits (location, age, mood, gender), learn conversational styles from past interactions and take actions using tools such as robotic process automation (RPA).
AI and the Customer Experience
Chris Radanovic, a conversational AI expert at LivePerson, said conversational AI can help consumers 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.
Conversational AI is a key for many brands that wish to improve the customer experience. Radanovic explained that consumers and brands are embracing conversational AI because it can provide personalized experiences that are quicker and more convenient than traditional ways of interacting with brands. “Think waiting on hold for a phone call or clicking through tons of pages to find the right info. Along with a more personalized experience, AI can also help to eliminate the pain points in the customer journey.”
Types of Conversational AI
Conversational AI is typically used two ways: actively and passively. It is used actively during communication between humans and machines and passively when it observes communication between humans.
Examples of active conversational AI include:
- Digital personal assistants: Virtual helpers like Alexa, Siri and Google Assistant.
- Digital customer assistants: Found on websites, built into smartphones and on apps to order services, like food delivery.
- Digital employee assistants: Allow employees to access information faster and streamline tasks.
Erik Duffield, GM of Deloitte Digital’s Experience Management Practice, said the sheer speed at which conversational AI and machine learning can operate is very effective at making decisions based on actionable data.
He noted the competitive ground in digital experiences has moved to a massive number of small decisions and interactions. “We are now seeing digital experiences shift from human to machine interactions, with AI and NLP enabling companies to execute their strategies at the speed and volume required to deliver the experiences that are expected by customers.”
Today’s conversational AI tools use predictive analytics to make decisions. Predictive analytics uses data, statistical algorithms and machine learning to discover the likelihood of future outcomes using historical data and statistical modeling. Conversational AI uses predictive analytics to determine the next “best step” in the customer or employee journey. Businesses can also use AI for fraud detection, managing resources and reducing risk.
Hospitality brands, such as restaurants and hotels, can use predictive analytics to determine the number of guests on any given night, which allows them to maximize occupancy and ROI. Retailers can use predictive analytics to forecast inventory requirements, configure the store layout to maximize sales and manage shipping. By analyzing past travel trends, airlines can more appropriately set ticket prices.
Machine Learning Has Existed for Years
Machine learning is nothing new, however. It’s just that the technology is moving forward rapidly.
“Machine learning programs and projects have existed in organizations for a number of years now, advancing from statistics and analytics toward data science. Out of that proving ground emerged a future goal (or requirement) that companies will not only have a few ML models, but dozens, if not hundreds, operationalized and embedded into consumer experiences,” said Duffield.
“This shift is leading to a new class of technology known as MLOps,” he continued, “which appears to be following the same maturity path as application development: Continuous Integration/Continuous Deployment, deployment automation, testing automation, etc. Leaders will build capabilities in this space and machine-driven decisions will be automated, validated, tested and measured.”
How Does Conversational AI Work?
Several components enable conversational AI to have human-like conversations through voice or chat:
- Automatic speech recognition (ASR)
- Natural language understanding (NLU)
- Dialog management
- Natural language generation (NLG)
- Text to speech (TTS)
Although AI enables applications to quickly make decisions based on actionable insights gathered from data, several steps are involved in the process. The initial step occurs when the AI application receives data from a human through either text or voice input. By using automatic speech recognition (ASR), the AI application can understand spoken words and translate them into text.
Natural Language Understanding
Next, the AI app determines what the text means by using natural language understanding (NLU) and then formulates a response to the text. The app accomplishes this by using dialog management, which creates an understandable response by using natural language generation (NLG). The app delivers the response as text, like an AI chatbot would do, or voice (using text to speech), such as Alexa would do.
Finally, the AI app uses machine learning to accept corrections and learn from each experience. This process enables it to produce better and more accurate responses in the future.
When creating conversational AI, it’s essential to consider the questions customers will have. Remember, they’re looking for relevant information and help solving a particular issue. In many cases, the essential ingredients to build conversational AI are already at hand: FAQs.
FAQs are the building blocks for developing conversational AI. They provide intelligent insights into the primary needs and concerns of users. Developers create FAQs with the information provided by the support team, as they’ve dealt with these questions repeatedly.
Once finished, the AI tool can answer common questions, freeing up time for customer support to focus on bigger issues.
Let’s take the example of a baggage FAQ for an airline. It might look like this:
- How many bags am I allowed to carry on?
- Are there exceptions to this allowance?
- Do baggage fees apply to all U.S. territories?
- What is the policy for active U.S. military personnel?
- If I packed my camera in a suitcase, will it be covered under your airline’s liability policies?
- What happens if my luggage is lost?
As you develop your conversational AI tool, you could add or remove questions as needed.
Understands Variations of FAQs
It’s essential to teach AI tools to recognize the different ways a customer might ask a question. For instance, the lost luggage question could be phrased three (or more) ways:
- How do I get reimbursed for lost luggage?
- How long does it take to find lost luggage?
- Who do I speak to about lost luggage?
An analytics team can provide the web data needed to understand context and the different ways customers might ask a question.
Next, build an “entity” (a segment of user input that describes useful information) around lost baggage information. It should contain values relevant to the category: lost, baggage, reimbursement, baggage claim number, destination and flight number.
Once assembled, these elements create a conversational AI tool that can have meaningful dialog with a curious customer. The end goal is an AI bot that satisfies customer inquiries and frees up human resources.
Benefits of Conversational AI
Building a conversational AI tool is good for a variety of business processes:
A conversational AI tool is particularly useful for small- to medium-size businesses that cannot staff an entire customer service department.
AI can deal with common customer questions and can operate 24 hours a day. With only a smaller support staff needed for critical issues, companies can save on training, salary and onboarding costs.
Providing Engagement That Increases Sales
Conversational AI tools provide customers with relevant information when it’s needed, regardless of the time of day.
As such, customer response times go down and satisfaction rates go up, increasing the likelihood of a purchase.
It is much more cost-efficient to scale conversational AI infrastructure than it is to hire and train new employees. Businesses can also use AI to save money when moving into new territories or dealing with seasonal demand spikes.
The Challenges of Conversational AI
Conversational AI applications rely on conversation data. Programmers train them on the use of a maximum likelihood estimation objective and/or reinforcement learning. Retraining is often required, even if there’s only been a small change in the conversation.
Data preparation and training can become an expensive endeavor. Additionally, conversation responses are based on business logic, which is industry-specific and challenging to describe. Decoding such logic using text data alone is practically impossible at this point. With voice-based conversational AI, many other challenges come into play.
Interpretation of Nonverbal Cues
When people speak to one another, voice itself is only one way they communicate. AI detects fluctuations in tone, hesitation and volume and interprets appropriately.
Other nonverbal cues, such as facial expressions, eye movements and hand gestures, are impossible for AI to detect unless the medium is video. As such, the importance of voice interpretation increases immensely.
Users’ Degree of Knowledge
Another challenge is the varying degree of knowledge of each person communicating with the AI bot. Children are limited in their knowledge and vocabulary and must be spoken to in an age-appropriate manner. Adults with different levels of education or experience in a given industry also must be spoken to with a response that is likely to be understood.
Location, Language, Sentiment
Location, language and sentiment also present problems. For instance, it’s difficult for AI to pick up speech when more than one person is speaking or there’s background noise. Understanding dialects, comprehending feelings and detecting attitudes, such as sarcasm, also bring up issues.
Not only must AI be able to understand conversation the way a human does, but it must also present complex or abundant information in a way that’s simple and easy to understand. But there’s no one-size-fits-all solution.
Other challenges AI applications face include privacy, security and user apprehension. Personal data has become a hot topic in recent years, with data infiltration attacks becoming more common. As such, people are reluctant to hand over information to brands.
Many people believe AI assistants like Alexa and Siri listen all the time. And with stories coming out about private conversations recorded by AI and used in the court of law, it’s not hard to see why some are wary.
Companies that choose to use AI applications must build strong privacy and security standards. Once developed, they should also make those standards known to users.
Related Article: Conversational AI: Creating a Framework of Ethics and Trust
Conversational AI enables people to use natural language to communicate with machines. You’ll find it in contact centers routing calls, in online chatbots answering customer questions, in automobiles assisting drivers and much more.
As AI applications progress and developers overcome challenges, we’ll likely see a future where AI conversations — even ones dictated by tone or body language — are indistinguishable from human interactions.