A bot is a computer program that automates a task. Chatbots are, therefore, programs that automate conversation, both voice and text. In essence, a chatbot can simulate an intelligent conversation with a human. They generally make use of machine learning to gather conversational rhythm, beat and flow, and thus mimic human conversation.
Chatbots are primarily used in customer service to shorten the response time when a customer contacts an organization. But the possibilities are endless. Chatbots can be used in many organizational activities and tasks including the following:
Research and development in chatbots grew after Facebook opened up their Messenger platform to make it easy for anyone to build a chatbot on top of their Wit.ai Bot Engine. This effectively made it easy to quickly and easily deploy a chatbot that interacts with customers.
The best way to understand the types of bots available is to examine them on a linear spectrum. On one end, you have chatbots designed to facilitate human conversation and on the other you have programmatic chatbots designed to completely replace human conversation. In the middle, you have various hybrids of either of the extremes. Most chatbots that lie in the middle of the spectrum have some kind of information-gathering feature. Essentially, they will collect information from the user and provide useful responses where possible. Where they are unable to do so, they redirect the user to a human being.
The latest generation of chatbots rely on artificial intelligence (AI). This allows them to learn and improve over time. Machine learning (ML) is at the heart of chatbot research and development. ML allows an application to input new information obtained from the user or even other sources, learn, and ultimately use that new knowledge to improve its future interactions. In essence, chatbots are now able to answer fairly complicated questions and not simply structured or scripted questions as was the case in the early days of development. The holy grail is to build chatbots so advanced that users aren’t sure if they are communicating with a computer or a human being.
The design and development of chatbots has so far been code-heavy. Chatbots can be created using any programming language with a web API. The most common deployments use Node.js and PHP in addition to several other libraries that support Python or Java.
The trend now is to move to chatbots that can be deployed with little or no coding involved. This enables faster delivery of applications and leads to cost savings. Applications are assembled and configured in a graphical interface. Think about the way assembly lines automate processes in auto factories. In the same manner, low code chatbot development automates difficult and time-consuming tasks.
However, no code deployments generally don’t offer much flexibility to developers. Thus, no code deployments are ideal for information gathering chatbots and those that facilitate human interaction, whereas low code chatbots are great where developers need to add custom functionality. The need for custom functionality should never be underestimated.
There are numerous tools available to build low code voice and text chatbots. Below are some of the tools which I consider worth exploring further.
Dialogflow comes with a web interface to create bots. This makes it a fairly easy low code solution for anyone to create basic chatbots. It integrates well with several common applications such as Google Assistant, Slack, Facebook Messenger, and Twitter and comes with basic in-built web integration. Dialogflow supports over 20 major languages and is free for basic use which is ideal for small businesses. The enterprise version costs $0.002 per request.
Amazon Lex comes with an easy-to-use web interface and runs on the same machine learning platform as Amazon’s Alexa. It integrates beautifully with SMS, Facebook Messenger, Slack, Twilio, Kik, and several other applications. Lex also has a basic chat user interface for testing on a website. At 10,000 free text requests and 5000 free voice requests per month in the first year, the service is worth a try for anyone looking to implement chatbots. The only drawback is that it is currently only provided in US English.
Azure offers a service that’s great for creating and publishing intelligent bots via a web interface. It integrates with Facebook Messenger, Slack, Skype, Kik, Telegram, Twilio, and many others. Azure Bot Service also comes with an Open Source web chat widget that’s available on Github for web and mobile integrations. The service supports multiple languages including English, Spanish, German, French, and many more. It is free for up to 10,000 messages per month, while paid plans start from $0.5 for 1,000 messages.
Wit.Ai has a user interface to set up intents and try it out. However, despite being a trailblazer in this field, the platform remains a bit too developer-centric. Non-techies may have a difficult time figuring it out. It has no direct integrations but comes with HTTP APIs and libraries available for Node.js, Python, Ruby, and Go. It also does not come with any web or mobile integrations. On the upside, being a Facebook service, it supports over 50 languages and is free for both personal and commercial use.
In addition to an easy-to-use interface to create a bot, IBM Watson comes with video tutorials and ready samples to help get your first chatbot up and running in no time. It integrates well with Facebook Messenger, Slack, WordPress, and Custom APIs. IBM Watson also offers a basic chat user interface for websites. It supports ten major languages and the free plan comes with 10,000 messages per month. Paid plans start from $0.0025 per message.
Botpress is an open-source platform hosted on Github that offers a graphical user interface where everything is clickable. You can create text- or pattern-based (e.g. keywords) bots. It integrates with Facebook Messenger, WordPress, Slack, Telegram, and many other applications and supports over 20 languages. While it is free, there is a paid version with premium support. For non-techies, this is probably one of the best low code chatbot solutions out there.
This list of low code chatbot solutions is by no means exhaustive. There are numerous other solutions I haven’t specifically included. However, if you are looking to implement a chatbot solution for your business, this is a good place to start.