September 1, 2021
A chatbot is a computer program that simulates a conversation with a human, usually through text exchanges and with the aim of answering queries. It’s now a key part of customer support for many businesses. Last November, Facebook revealed it hosts 34,000 chatbots through its Messenger tool.
Why Are Chatbots Getting More Popular?
• Chatbots are rapidly improving, making them a genuinely useful tool rather than a gimmick.
• Businesses expanding online means they risk growing an audience or customer base more quickly than they can expand their human customer service workforce.
• Because chatbots work 24/7, they’re a useful alternative for companies wanting to provide out-of-hours customer service that would otherwise have to choose between expensive overtime and unsocial hours payments to domestic staff or potentially inadequate contracting out to overseas agencies.
• Chatbots are getting more mainstream because it’s possible to build them into existing chat technologies with which users are familiar, such as Skype and Facebook Messenger.
• The inclusion of “digital assistants” such as Siri on iPhones and Google Assistant, along with voice control and recognition devices such as Amazon Alexa and Google Home, means people are becoming comfortable with the idea of interacting with automated systems
What Types of Chatbots Exist?
The jargon used around chatbots can be confusing, with different terms used in different ways. As a rule of thumb, you can categorize a chatbot by answering two questions:
1. Does it use artificial intelligence?
2. If so, which AI elements does it use?
A chatbot that doesn’t use artificial intelligence is often called “rule-based.” It’s more about being fast and reliable than clever. The rules it follows effectively come down to “If the question mentions this topic, then give this answer.” In effect, the chatbot acts as an automated Frequently Asked Questions page. An artificial intelligence bot is one that replicates some of the benefits of using human staff, usually with Machine Learning and/or Natural Language Processing.
Machine Learning means the chatbot learns from its experiences and becomes more useful and intelligent. For example, if it interprets a question one way and the site visitor then has to restate the question, the chatbot learns a new link between a particular input and the desired response Natural Language Processing is effectively the way a chatbot deals with the fact that humans don’t always phrase a question in the most consistent or logical manner, and that replying to the user in a more natural and conversational way will give them a better experience that feels more personal and responsive
Where Should I Start With Chatbots?
Unlike many tech fields where a few companies and services dominate, the chatbot world is still filled with dozens of alternatives. Two platforms that might be a good starting point are:
• Facebook’s Bots for Messenger: Great for integrating chatbots into a familiar setting without needing to learn too much new technology. For example, SnapTravel lets users search for unsold hotel rooms by answering a few simple questions in chat format rather than filling in a web form.
• IBM Watson Conversation: Based on the system that was smart enough to win Jeopardy!, this chatbot service took the entirety of Wikipedia as its starting database. You’ll need to pay to get the most from it, and you’ll definitely need to plan ahead to make the most of it. Staples found a particularly creative way to use the technology: a physical button with a microphone that lets customers order office stationary or check up on orders.
Others you could consider include Chatfuel, Pandorabots and ChattyPeople. It is easy to get carried away with tech and getting features you don’t really need. Before signing up to a service take a little time out to first work out exactly where your business needs to improve and then explore how chatbots could help that process