How do chatbots work? An introductory guide
Wondering how chatbots work?
Put simply, chatbots follow three simple steps: understand, act and respond. In the first step, the chatbot processes what the user sends. Then, it acts according to a series of algorithms that interpret what the user said. And finally, it picks from a series of appropriate responses. Chatbots can be programed to respond the same way each time or to respond differently to messages containing certain keywords.
Chatbots are deployed to mimic real, human conversations. You're likely to have encountered one of these automatons online or through a third party messaging platform like Facebook Messenger or Slack. Chatbots rely on natural language processing, the same technology behind virtual assistants such as Google Now or Apple’s Siri.
Why are chatbots important for customer experience?
Chatbots enable businesses to automate aspects of customer service without tasking customer service representatives or salespersons. According to Forbes, 60% of Millennials have used chatbots when shopping, and 70% of them reported positive experiences. Many businesses make chatbots available for their customers when they have a question or run into a problem, but some businesses are leveraging chatbot technology even more. Estee Lauder utilizes a chatbot embedded in Facebook Messenger that uses facial recognition to pick out the right shade of foundation for its customers, and Airbnb has used Amazon Alexa to welcome guests and introduce them to local attractions and restaurants.
Chatbots are useful for conversational purposes when an app can't perform an action because multiple variable inputs are needed to solve the problem. In other cases, a chatbot will offer simplicity by providing the most immediate and direct solution to a person's question.
What is the current state of chatbot technology?
If you asked somebody years ago about chatbots, the response would have been hyperbolic. Apps’ retention rates were slowing down, while messaging apps were on the rise. It was not uncommon to see articles with headlines like "This is how chatbots will kill 99% of apps."
And there was data to back up the hype. Facebook Messenger crossed the 1 billion users mark in 2016, the same year it introduced its Bots API, which boasts over 100,000 bots. Western app developers were also enamored by apps like China’s WeChat, a much more immersive messaging platform whose bots are capable of a myriad of tasks, like calling a taxi or making appointments.
However, being excited about chatbots and actually implementing a working chatbot are two very different things. In order for users to engage with a chatbot consistently, it needs to be easy to use, efficient, and get the job done. Developers and investors frequently run into snags when launching a chatbot — but we have the expertise to help you get things right.
How chatbots work
Any system or application that relies upon a machine’s ability to parse human speech is likely to struggle with its complexities, such as metaphors and similes. Because chatbots rely so heavily on natural language processing, they are also constrained by its limitations. A user will only tolerate a chatbot if it speeds up communication and makes things easier.
The other issue is human speech itself. Chatbots try to mimic conversations, but most conversations aren’t linear: discussions restart, there are tangential topics, or multiple topics being discussed at once. This can be very tough to follow algorithmically.
One of the best ways to improve a chatbot is to constantly train it and get it to respond to different human interactions. The more data you feed a chatbot, the better it can adapt to human speech and all its idiosyncrasies, and the better you get at achieving a natural, human-like conversation. Successful chatbots rely on context understanding, intent variation and intent recognition.
Context understanding for chatbot training
Context understanding is the ability to remember and track different aspects of a conversation including location, time and preferences, and then to combine all of the inputs to paint a picture of the conversation. Just like humans use surrounding context to inform their interactions, chatbots also need contextual information to maintain an effective conversation.
Intent variation for chatbot training
To understand intent variation, it’s important to first understand what intents are in natural language processing. An “intent” refers to the purpose of a user’s input into a model, such as a chatbot or search engine. For example, if a restaurant has a chatbot on their website, then a customer might use it to ask about business hours.
This is where intent variation comes in, because people have different ways of expressing the same intent. The customer might phrase their question as “What are your business hours?,” “What are your hours of operation?,” or “What time does your restaurant open and close?” The chatbot should be able to understand that there are different ways of expressing the same intent, and inform the customer of the store’s business hours regardless of which way the question is phrased.
Intent recognition for chatbot training
Intent recognition is the ability to extract relevant information from each word and sentence, and understand the intention and the meaning behind it. This allows the use of long complex sentences by users, because the chatbot is able to understand and extract multiple intents.
These complex features require a robust natural language processing foundation that can only be sustained by high-quality training data. That’s where TELUS International comes in. We can provide high-quality training data and everything you need to deploy an efficient bot. Start a conversation with us today.