What are chatbots?
A chatbot is a software application that uses artificial intelligence to conduct a conversation using either text or speech.
The are a few different types of chatbot
- Script-based chatbots are the most basic type of chatbot and are basically decision tree hierarchies presented presented as a selection of prompted answers from a menu or a button.
- Keyword based chatbots reply with the best response based on keywords in the questions.
- Contextual chatbots are by far the most advanced of the three bots discussed in this post. These chatbots utilize Machine Learning (ML) and Artificial Intelligence (AI) to remember conversations with specific users to learn and grow over time.
It may feel like chatbots are relatively new, but they were first created in the 1960s.
After the Turing Test was created in the 1950s (by Alan Turing) the race was on to try and become the first engineer to design a machine that could beat the test. The test was to see if a computer program could convincingly impersonate a human in a real-time conversation (written or oral) with a human judge.
The first program credited with a victory was ELIZA (named after Eliza Doolittle, the working-class hero in George Bernard Shaw’s Pygmalion) created in the MIT Artificial Intelligence Laboratory by Joseph Weizenbaum between 1964 and 1966.
Fast forward to today and chatbots are now being used commercially.
They are sometimes also referred to as a conversational agent, chatterbot, interactive bot, conversational interface, conversational AI or talkbot.
Where have I seen one of these before?
If you’ve ever jumped onto a brand’s Facebook Messenger to ask a question and are immediately greeted with a salutation and some qualifying questions, you were probably dealing with a chatbot. There are over 350,000 active bots on Messenger.
The Wall Street Journal (click on Send Message) have a great chatbot… it might not pass a Turing Test but it certainly presents information in a fast and consumable way.
It’s not just Messenger that can host chatbots, of course. Lots of messaging platforms now have the capability to use chatbots. Another great example is the Irish company, Intercom who have a chatbot both on and off their website as part of their customer service platform.
Is it hard to make a chatbot?
This technology has become both increasingly popular and easier to deploy.
The two examples above use technology that is built-in to the platform.
There are some standalone apps you could also use. Or if you have the budget or the technical staff you could build your own.
Why are chatbots useful for business?
In 2018, every minute there were
- 18 million text messages
- 38 million WhatsApp messages
- 25,000 GIFs sent via Facebook Messenger
Piggybacking on this rise in popularity, chatbots have been adopted most commonly in a customer service function for businesses.
If the product or service is straightforward the chatbot can provide a lot of information within the conversation timeline (like The Wall Street Journal example). If your product has many variations or is more consultative in nature, chatbots can be used to qualify what type of query the customer has.
For example, an insurance company might be able to screen in customer inquiries in the following way.
Each question informs the next question based on the answer given. You can create a decision tree based on the various question and answer permutations.
If a customer service representative does need to intervene you can make sure they understand the nature of the query and can be prepared with the information to be able to provide answers.
How can I find out more about chatbots?
Marketing superpower HubSpot have created this impressive article that brilliantly makes the case for chatbots.
This article will help you set up a Facebook Messenger chatbot reasonably quickly.
If you want to delve a little further Maruti Techlabs have created this deep dive article into all things chatbot.