Chatbot Development Platform Vs Chatbot Publishing Platform.
Along with the traditional methods, Mobile application & website for user interaction, now businesses have another method for user communication and engagement. This another new method is chatbot. Chat Bots mimic humans and provide the response to user basis on predefined rules or from knowledge base using artificial intelligence. There are different channels for chatbots viz. Facebook messenger, Line, Skype, WeChat, iMessage etc. These channels offer developers with the APIs and platform for both development and deployment. Although different channels have different set APIs and protocols for integration, the end result is almost the same, i.e. to answer the user query. Since every business doesn’t have the technical expertise to consume these APIs offered by chatbots channels, so this has lead to the formation of third-party Platforms which facilitate end user with chatbot development and deployment using simplified user interface. Coming into existence of such platforms, a confusion has arisen which ones are the actual developments platforms. To understand it in the better way we can divide these into two categories Chatbot development platforms and Chatbot publishing platforms.
Chatbot Development Platform
These are the third party platforms which consume APIs offered by chatbot channels Facebook messenger, Skype, Line, Kik etc. and provided the end user a simple graphical user interface for bot development. Using these platform once need not to have technical expertise and underlying artificial intelligence engines. The user just needs to define their bot flow I and configure it in the development platform. Examples of chatbot Development platforms are :
These development platforms are either cater to all service areas with few constraints or few are dedicatedly working to carer the development in the specific domain. E.h Foodbot.ai is a bot development platform which is fully focussed on Bots developments for the restaurant industry and covers almost every small segment of this industry with comprehensive reporting.
Chatbot development platforms must comply with the chat publishing platforms policies in order to run the platform smoothly.
Chatbot Publishing Platform
Chatbots publishing platforms are the chatbots channels, who have already instant messenger platform used by millions of their subscriber. These publishing platforms offer the development platform APIs for bot developments and tools for publishing the bots on their messenger platform. Once the bot is published, users may discover the bot and start interacting with the respective business service.
Examples of Chatbot publishing platforms are:
- Facebook Messenger
Chatbots can be programmed in two ways, rule-based and smart machine learning based. These two types of bots can be divided into three generations. An understanding of the different bot generation will help you to decide which bot best suits your business needs.
Three generations of bots are :
- first Generation Chatbots
- Second Generation Chatbots
- Third Generation Chatbots
1) First Generation Chatbots
First generation bots use the very simple way of bots implementation. These are rule-based and have predefined rules and response to the user inputs. First generation bots do not use artificial intelligence. Their response is very predictive.
Example: in a restaurant but we can define a rule for the menu, address etc.. When the user interacts with the bot, it gives options like Menu, Address. The bot will check the rule defined for each and display the respective item to the user.
2) Second Generation Chatbots
Second generation chatbots make use of artificial intelligence while responding to the user queries. 2nd and 3rd generation chatbots conduct Natural language processing know as ‘NLP’ while responding to the user.
There are two approaches to the NLP, Supervised and unsupervised machine learning. This variance in the NLP implementation separates the 2nd and 3rd generation bots. Second generation bots use supervised machine learning. In supervised machine learning, there is the mapping between various input patterns to the output response. If any of the patterns match then the prespecified response is given to the user.
For ‘Rain boots’ we can define various patterns like Rain Boots, Gumboots, Wellies, Galoshes etc.
3) Third Generation Chatbots
Third generation chatbots are similar to 2nd generation chatbots with the difference that they make use of unsupervised approach of machine learning along with the supervised machine learning. There is no labeling and mapping of user input patterns to the output. In unsupervised machine learning, Artificial intelligence system learns about the regularities in the data by modeling the underlying structure in the data.