Dialogflow

The power of Natural Language Processing (NLP) is now available for everyone to use. DialogFlow basically enables us to create our own machine learning model for AI conversations without coding. Guess what! We can do it for free.

You just need to create a few intents or questions that states the user's intention. This could be done by just typing in some possible questions users might ask for a particular case and adding a few replies for that. After training the model you can integrate it with other application like Facebook Messenger, Telegram, Twitter, Slack or Skype for example or it could be integrated on your own website (an e-commerce website for example to sell your products).

You can also build applications on top of Google Assistant. In fact Google Assistant uses the same DialogFlow engine to converse with you. Another interesting feature is that the AI agent you created can also be integrated in to your IoT projects (like switching ON and OFF a light bulb), but it needs a bit of coding knowledge.

DialogFlow is available in many languages like German, Spanish, Hindi and a few other languages too. This list is increasing periodically.

There is an enterprise edition of DialogFlow called DialogFlow CX or Customer Experience with some added features for enterprises. Also there are some limitations on the free version but it won't cost too much for the upgrade and the cost is based on usage like any other Google cloud feature.

Some of the features in DialogFlow are still in beta phase. The number of available languages as of now is limited. It would be better if some more languages were included. Most of the advanced features in DialogFlow might be difficult to understand for regular users. It is better if you have some programming background.

What I liked most about Dialogflow is how easy it is to create chatbots and understand user conversations, even without much coding.

What I liked least about Dialogflow is that it can get complicated when trying to handle more advanced features or integrate with other systems.

I appreciate Dailogflow software's one-click integration and context management features. I used them to deploy chatbot across multiple websites and mobile apps. Plus, it helped me and my development team customize its responses to our customers questions. Additionally, Dialogflow's project testing console and natural language understanding capabilities ensured our automated chatbot responses were relevant to intent of customers' questions. Also, it kept in mind related questions which would be asked in the near future by the same customers. It helped implement artificial intelligence and machine learning.

The design of web chat integration is not very compelling. Also, our developers thought Dialogflow documentation around Python SDK is limited in scope. it has limited examples of implementation, use cases are not diverse and few tutorials on how to implement. As an admin of Dialogflow, it makes it complex to configure various actions or create custom components of the chatbot application. In a nutshell, using Dialogflow, it often becomes challenging to customize chatbot flow when a marketer has limited technical expertise.