In the long run, the world will evolve in computing from a mobile-first to an AI-first approach. - Sundar Pichai, CEO, Google Inc.
While it is quite easy to dream about a "machine first" future, we still know only the tip of the iceberg. Not long ago, people used to have a good laugh about chatbots for their lack of reasoning. After all, many people had no idea what a chatbot was, and little did they know about how Eliza (the first Chatbot) or Parry (the chatbot which simulated a patient with schizophrenia) worked. To them, it was all about a lady at the bottom of a page, giving out dumb responses for their questions.
Over the years, the boom in chatbot and machine learning applications has led to businesses crafting their online presence in the form of websites and using artificial intelligence for a better customer experience. This is not surprising, given the fact that in recent years, chat or messaging has taken over social media to be the "go-to" option for users who want to contact a business.
Technological advancements have provided a new approach to real-time, personalized customer experience by implementing deep learning algorithms, particularly by feeding large data into the bots. However, content marketers over the years have been working to make content behave more like a chatbot. Every improvement in the form of automation, personalization and customization is concentrated towards providing an engaging, tailored-to-specific-needs content for a user.
While many marketers still have a different opinion of bot development being something outside the norms of traditional CMS, businesses have realized the transformative value of chatbots and have started treating them like another channel. A channel to allow continuous flow of content by integrating them with the CMS.
One such CMS that allows seamless integration is Drupal. Its ability to be used as a decoupled CMS makes it the best choice to serve the content of a chatbot's response.
Drupal 8's chatbot API facilitates this implementation of chatbot and surfaces content without having to write thousands of lines of code for every single AI you use.
Chatbots are no longer the old clunky machines that they once were. The high-end technology has improved the way they work and has allowed them to provide excellent service by solving customer issues, book flights, and even order pizzas. A report by Gartner studied the developments in chatbot usage among businesses and according to the report, 20% of the business content could be machine generated by the end of 2018.
With seamless user experience as a predetermined goal, businesses use chatbots to interact quickly with their audience and make it feel personal and real-time. Thus, the intent of the user is an important factor that they focus on. It defines the range of possible responses that a chatbot can come up with for what the user is looking for.
For example, if the user is looking for an information on movies, he might type:
How do I book a flight to Delhi?
By mapping the exact phrases to specific intents, the AI is able to determine what the user is looking for.
For example, the intent is : How do I fix my Television?
In both of the above cases, mapping can be done using entities such as,
- How do I book a @means of transport to @place in the former's case, with the entities being the various modes of transports and places fed to the bot
- How do I @task my @appliance in the latter's case, with the entities being different tasks and list of appliances.
Chatbot API - What does it do?
Drupal as a "headless CMS" should allow you to surface content through the various available AI services such as Google Home, Amazon Echo or other services. Before the chatbot API, each intent required a custom code on each of the chatbots to be used.
With Chatbot API, it provides a common layer such that you only need to write your custom code for each of your intent once. However, the Chatbot API by itself does not take care of anything and should be installed only when another module asks for it. These specific modules handle requests and responses on the bot.