Dialogflow Features

• Developer collaboration: You can invite developers and other team members to collaborate on your Dialogflow conversational agent(s).
• Conversational IVR (interactive voice response): In addition to NLP (natural language processing), Dialogflow uses voice-driven, hands-free customer self-service that uses NLU (natural language understanding) to understand the content and context of spoken requests. The intent of this technology is to remove the burden on customers to navigate through slow, confusing and hierarchical menus.
• Prebuilt Agents: Dialogflow’s pre-built agents are customizable and specialize in different areas of knowledge. If you find an agent you want to use from Dialogflow’s menu of pre-built agents, you can import them into your account and set up a webhook to get the agent responding. You’ll have access to all the parts of the agent, so you can modify or reuse intents and entities in your projects.

• Intents: Intents are basic building blocks of conversation in Dialogflow. Based upon the voice/text input of users, Dialogflow selects the best matching intent and replies back with a response. This works by using “training phrases.” Then, with intent, you can define training phrases which will be matched with user inputs. As Dialogflow finds a matching training phrase, it will call upon corresponding intent and reply back to the user with a response set for that intent.
• Analytics: Once your chatbot is live, you can see how your chatbot is performing. Dialogflow Analytics provides insights into the usage of your conversational interface. By using these analytics, you can try to find out more about user behavior and make improvements to your chatbot.

• Entities: Entities are Dialogflow’s tool for identifying and extracting useful data from natural language inputs. While Intents allow your conversational agent to understand the motivation (or intent) behind a specific user input, Entities are used to select specific pieces of information that your users mention. Dialogflow offers a small number of pre-built system Entities for you to choose from as well.
• Fulfillment: While Intents are the way to design responses for the end-user, which might be able to handle all the responses of your bot when responses need to be dynamic, you’ll need to get the data from your server. It’s these scenarios that give the Fulfillment feature its reason for existence. Dialogflow provides webhook integrations that enable communication of your chatbot with your server, and if you want to deliver a response from your database, then you can use a webhook to “fulfill” that response.
• Small Talk: Dialogflow’s conversational agents have the capability to learn how to support small talk without additional development. In these cases, your bot can automate responses with predefined phrases, set by you.

Dialogflow Pros and Cons

Pros - Dialogflow

• Build conversational interfaces for multiple platforms: With Dialogflow you can create bots and apps for Google Assistant, Cortana, Alexa, Facebook Messenger, and other platforms. Additionally, you can engage with users through phones, wearables, cars, speakers and other smart IoT devices.
• Scalability: Because Google’s resources are so massive, Dialogflow has the ability to scale. Additionally, Google Cloud offers a server-less way to run your conversational agent on the backend.
• Integrations on major platforms: Dialogflow offers integration with a range of chat platforms like Google Assistant, Facebook Messenger, Line, Slack, Telegram, Skype, Viber, Twitter, Twilio, Cortana, Cisco Spark, and Amazon Alexa. Once your bot is developed on Dialogflow (with some modifications required) it can be integrated on multiple platforms. However, the only thing that makes this unique is the available platforms to build on. The difficulty to do so, however, actually makes this a less favorable process than what’s available on other, better AI-powered chatbot platforms.
• Multilingual: Dialogflow has 20+ supported languages including Spanish, French, and Japanese. Again, especially for Google, this is not all that impressive, as there are many other fantastic chatbot builders available with more languages available.
• Supported programming languages: Dialogflow provides client libraries and guides for C#, Go, Java, Node.js, PHP, Python, and Ruby.

Cons - Dialogflow

• Flexibility and developer experience: Although Dialogflow is fairly intuitive on the surface, you’ll find that it’s not as flexible a platform as you would have hoped for. For example, if you decide that I want to move a follow-up response under a different Intent, you cannot simply drag that under the desired Intent. Instead, you’ll need to delete the existing Intent, create a new Intent in a different location, and re-type all of the training phrases you’ve already created. This poses a problem because it results in a lot of tedious repetitions which are quite frankly a waste of time and it forces developers to think well in advance about the hierarchical dialog flow (no pun intended) of their conversation. This means there’s limited flexibility to change things you create in the future, which needless to say is a major issue.
• Limited webhooks and integrations available: You can only provide one webhook for each project. This essentially means that the entire chatbot must have exactly one webhook instead of choosing multiple webhooks on an intent-by-intent basis.
• Customer support: Google isn’t too big on live customer support, nor is Dialogflow a very notable customer support chatbot option. The only support you’ll get from Dialogflow is from filling out a support ticket online and waiting to hear back. Even on the Enterprise version of the platform, the only additional support you’ll receive is a guaranteed support ticket response time.

• Lots of manual work, time and training: In many instances, Dialogflow makes it harder than it should be to automate processes and expand your conversational agent’s learning. This can get annoying because you have to input many things manually, especially when you consider the need to train your bot over time. Therefore, the ability to even make your chatbot better is hindered, which unfortunately defeats one of the platform’s main purposes.

Again, Dialogflow is not built for businesses looking to make quick improvements, and it’s not exactly ready for scale either. Dialogflow is for developers only, and it’s more experimental at this time than it is useful in the wild.

Dialogflow Review - Final Analysis

Dialogflow is designed by engineers and built for software engineers. Although Dialogflow is a cool option for programmers who want to explore NLP, the platform itself is still very much a work in progress and doesn’t provide for complex conversational user interfaces that have a lot of moving parts to consider.

So, even for software engineers, Dialogflow simply isn’t as robust a platform as many of the other top chatbot builders out there. The mapping of entities and intents is extremely complicated, there’s a limited number of webhooks to use for each conversational bot, and for a Google platform, the flexibility and developer experience is actually sub-par.

Ultimately, Dialogflow has some promise for the future, but as for now, the number of resources and the amount of effort required to create a business-ready, AI-powered chatbot is too great. Dialogflow is currently only useful for companies looking to experiment and have money to burn, or for developers with an AI-chatbot building hobby.