TLDR: Simple chatbots use rule-based processes to match keywords to information and respond appropriately. More complex approaches use machine learning to recognise user intent and continuously get smarter.
People are now spending more time using messaging applications on their phone than anything else, so it makes sense to be where your customers are.
The user experience
It’s simple. A site visitor clicks on the chat window on your website and types out a query. Behind each query lies intent, they may want to know some information, make a purchase, or use a service.
The bot is programmed to follow a process to give the correct answer. The goal is to capture the semantics and core meaning of the words sent to it, and to use that information to give the correct answer.
The technical bit
How exactly you create a chatbot depends on the software you use. There are many different softwares out there to choose from, or you could create a simple bot through Facebook messenger. Chatbots essentially work one of two ways:
Rule-based – these bots are programmed to recognise words and phrases, and response with a pre-written response
AI & Machine Learning – these bots are more complex, and can continuously train themselves as they handle more chats
Depending on your needs, both options could be a good idea. For simple FAQs, a rule-based chatbot could work exceptionally well, but for more complex conversations, using AI can have great results.
Rule-based chatbots are the most simple to train, and they use something called pattern matching to respond to queries.
The programmer will analyse a large amount of data, from previous conversation logs if there are some available, emails, and phone calls in order to compile the most commonly asked questions. They will then create a response and match up commonly used phrases and words to that response.
The bot will use pattern matching to match the text to the appropriate response. This works super well for businesses who are overloaded with FAQs, but it means the bot cannot go beyond this associated pattern.
Chatbots can use machine learning and deep learning to improve their intelligence, and the two main ways are Natural Language Understanding, and the more complex Natural Language Processing. Machine learning allows the bot to be retrained through feedback loops, provide A, B and C options that the customer can choose from, improving accuracy.
Natural language understanding
The bot is programmed using entity, intent and context.
Entity – an idea or concept is given to the bot, which is the entity. This could be ‘pizza’.
Intent – the intent is what the use is implying, and what action the bot should perform. E.g. if the user asks ‘do you do pepperoni pizza?’ the user is wanting to order a pizza.
Context – the NLU won’t remember the question it has asked once it is sent, so a state of phrases is stored to ensure the bot will retain the concept of the conversation.
Natural language processing
Natural language processing takes things to a new level and enables the chatbot to continuously get smarter.
Tokenisation – words are separated into different pieces (tokens) which are useful for the bot.
Sentiment analysis – the bot learns if the customer is having a positive or negative experience and whether they need to be transferred to a human.
Normalisation – the bot processes the text to check for common spelling mistakes, errors and typos.
Named entity recognition – the chatbot looks for categories of words such as the product name, customer information, whatever is required.
Dependency parsing – the boy looks for subjects and objects such as verbs, nouns and common phrases to find related phrases they may be trying to convey.
Customer service – bots work very well for customer service purposes. They can be trained to handle a vast number of FAQs, and integrations allow users to check order status and make informational changes with ease.
Lead generation – chatbots are just as effective when used for marketing purposes. You can train the bot to qualify leads, and integrations can enable users to book meetings or sign up for things.
Sales – chatbots can be trained to offer product suggestions, guide the customer through their buying journey and reduce basket abandonment.
Advertising – many companies also choose to use Facebook messenger for this purpose, as well as their own website. You can send marketing information, competitions, coupons and interactive activities to stay front of mind.
Technology is moving swiftly, and although bots are not yet perfect, the improvements in recent years have made them a great option for a range of business needs. Whether you need something to handle customer service queries or for marketing purposes, a chatbot can help.