Why do we use DialogFlow in the first place?
Google Dialogflow is natively a natural conversation understanding interface which provides support to many of your applications requiring human interaction. It enables configuration of automated conversations about various topics associated with your domain of interest. This has been considered as a convenient method for deploying chatbots without actually having to go through the whole procedure of training sophisticated models to enable machine learning. This also works with audio input recognition and understanding, responding in text as well as audio output based on your requirements.
What functionality is offered?
Google Dialogflow offers users the ease of creating a automated conversation assistant that can be set to answer various questions for users without actually having any coding knowledge. It is safe to say that even “noobs” of programming can create chatbots using Dialogflow, this also allows integration of various services upon requirement. A few note worthy integrations are, Twilio, Slack, Facebook Messenger, etc. A non-tech small and medium business can integrate chatbots on their website for free. Dialogflow is a platform that enables the development of chatbot without the need for installation, technical infrastructure, and coding knowledge. It is easy to integrate in many different platforms directly without extra coding. Hence both non-tech and tech people can easily develop chatbots for their purpose.
So why exactly do we have a problem using Dialogflow?
Thisb questions has many answers depending upon the users and their usage of Dialogflow. In this article I will try to cover all the points about why I faced problems with Dialogflow and would prefer models built by hardcoding gives us much more functionality and freedom.
- Lack of flexibility due to constraints of using API’s in Dialogflow environment. What this essentially means is that when we try to integrate a chatbot created on Google Dialogflow with various other applications, depending on the type of application the functionality in the chatbot is restricted, which makes us clearly furstrated, the lack of resources on specific integrations just add fuel to fire.
- Limited scope of scaling and inability to provide work on complicated answer systems where the user’s answers can be misclassifed into more than one intent depending on the nature of the response is also a great drawback.
- Restrictions in customizing various features is another drawback, Small changes which can be made easily just by adding a few lines of code cannot be done, just because the sourcecode of Dialogflow does not allow editing.
- Making a statement and trying to convey the message of your organization or expression of your theme is just not possible with dialogflow as customizaiton is hindered.
- The last point is that, it is impossible to block the matching of an intent if a context is present.
The are my reasons why I would not choose Google Dialogflow for developing a enterprise level application with Natural Language Processing capabilities. My reasons being as stated above, perhaps it never was created for that purpose, It actually empowers all those people from non-technical background into designing a chatbot which can actually be integrated to their website or application for free.