Dialogflow Review

By Joren Wouters Updated on

Dialogflow

Dialogflow is the natural language understanding platform of Google that you can use to create omnichannel chatbots.

But should you use Dialogflow to create your chatbots?

Let’s dive in!

 

What is Dialogflow?

Dialogflow is the natural language understanding platform of Google that allows you to understand language with your chatbot.

What Dialogflow tries to, is understand the intent of the user based on training sentences that you give.

Based on those training sentences, Dialogflow knows how to automatically respond to questions of your users.

Before we dive into the rest of this review, I wanted to let you know that it is also possible to integrate other chatbot builders with Dialogflow, so you can use Dialogflow for understanding the intent of the user and the other chatbot builder to handle the rest of the conversation.

 

User Interface

Dialogflow has two types of editions: Dialogflow Essentials (ES) and a Dialogflow Customer Experience (CX).

Both types of editions come with different interfaces and can be used for different purposes.

Dialogflow Essentials (ES)

Dialogflow Essentials is the standard edition of Dialogflow and most people are familiar with this one.

Within Dialogflow ES, you can understand the intent of the user by providing training sentences to your agent. The agent is then trained to understand these sentences:

dialogflow-intent

And create a reply for it:

dialogflow-intent-response

Also, you can use Entities with Dialogflow. With Entities, your chatbot can understand the difference between a “veggie pizza” and a “pizza”. In this example, the entity part is “veggie” that relate to a topping of a pizza:

dialogflow-entity

If you want more information on how you can use intents and entities in Dialogflow, check this video for intents and this video for entities.

Moreover, you can test your chatbot in the Dialogflow interface:

dialogflow-test-chatbot

 

Dialogflow Customer Experience (CX)

Dialogflow CX is a relatively new version of Dialogflow and is not yet used by many people.

In general, Dialogflow CX is not recommended for many use cases, unless you have a very sophisticated chatbot.

The benefit of Dialogflow CX is that you have a visual flow builder that provides a better overview of the conversation:

dialogflow-cx-flow-builder

But at the same time, I wouldn’t recommend using Dialogflow CX if you are a beginner with chatbots. The interface is much harder to use than the simple interface of Dialogflow ES.

In general, I think the interface of Dialogflow ES is easy to use and it is easy to set up your chatbot. Dialogflow CX provides a better overview, but also has a steeper learning curve (and you probably don’t need to use it).

  • Visual Flow Builder
    Create chatbot dialogs with a visual flow builder.
  • Test chatbot
    Is it possible to test the chatbot before putting your new flows or changes live?
  • Interface easy to use?
    Is the chatbot platform easy-to-use?
  • Easy to setup a chatbot?
    Is it easy to set up your chatbot with this chatbot platform?

Chatbot elements

Dialogflow provides enough elements to build a good chatbot.

They also lack some important elements, such as Delays, Video and Attachments.

Also, some chatbot elements are not built-in, but you can code these yourself using the Custom Payload feature.

  • Delays
    Use delays between messages in your chatbot
  • Images
    Can you send images via your chatbot?
  • Video
    Can you send video via your chatbot?
  • Audio
    Can you send audio via your chatbot?
  • Attachments
    Can you send attachments via your chatbot?
  • Gallery (Horizontal list)
    Show a gallery of cards inside your chatbot. A card consists of an image, title, subtitle and button.
  • Vertical list
    Show a vertical list of items inside your chatbot
  • Emojis
    Can you use emojis inside your chatbot?
  • Persistent Menu
    Can you add a persistent menu to your chatbot?
  • Quick Replies
    Can you add quick replies to your chatbot where users can click on?
  • Buttons
    Can you add buttons to your chatbot where users can click on?
  • Free text input
    Can chatbot users freely type text to your chatbot whenever they want to?
  • Location sharing
    Can users share their location inside the chatbot?
  • Email validation
    When a chatbot user gives their email, can the platform validate it's a correct email address?
  • Phone number validation
    When a chatbot user gives their phone number, can the platform validate it's a correct phone number?
  • Date validation
    When a chatbot user gives a date, can the platform validate it's a correct date?

Users

In Dialogflow, you cannot see a complete user overview of all the people that talked with your chatbot.

In fact, if you want to store user information, you need to create a custom integration with a Google Datastore or other kind of database.

  • Custom fields
    Store information from your chatbot users inside fields. For example, you can ask for someone's email and store it inside the "email" custom field.
  • Tags
    Add tags to users in your chatbot flows to organize and categorize contacts.
  • User segments
    Can you create user segments based on custom fields and tags?

Messages

With Dialogflow, you cannot send broadcasts and sequences. In essence, Dialogflow is user-triggered, which means that it responds when the user asks something, but the chatbot cannot initiate the conversation (or you need to code this yourself).

Also, with Dialogflow you cannot use conditions in your chatbot flows. It is possible to have different branches in your conversation by using Context, but this is kind of a workaround. If you want to use conditions based on user information, you probably need to code this yourself.

Furthermore, with Dialogflow it is possible to have a chatbot that can understand multiple languages at the same time. Unfortunately, it is not easy to translate these messages, because you need to manually add the same intents and same training phrases (but then in a different language).

  • Broadcasts
    Send a message to all or a part of your chatbot users at a specific date and time
  • Sequences
    Send messages on a pre-defined schedule with a certain time interval between those messages
  • Conditions
    Make checks in your chatbot flow based on if-then logic. For example, if the user has not given their email, ask for their email.
  • Multiple languages
    Is it possible to create a chatbot in multiple languages?
  • Easy to translate
    Can you translate text to other languages without using a third-party platform and without creating duplicate flows?

AI & NLP

One of the reasons so many people use Dialogflow, is because of their AI & NLP, which is really good.

As described above, you can train your Dialogflow agent to understand the intent of the user by providing training phrases.

Also, you can use Entities so your chatbot can understand useful information that your users say.

  • Keywords
    Trigger chatbot flow based on certain words or phrases a user says to the chatbot.
  • Intents
    Trigger a chatbot flow by understanding the intent of the user.
  • Entities
    Capture specific information from sentences. For example, if a user says "veggie pizza", the chatbot can understand it's not a regular pizza, but a veggie pizza.
  • Dialogflow integration
    Does the chatbot platform provide a direct integration with Google's AI Platform?

Channels

With Dialogflow, you can put your chatbot on 14 channels:

  • Your website
  • Facebook Messenger
  • WhatsApp (using Twilio)
  • Telegram
  • Twitter
  • SMS (using Twilio)
  • Slack
  • Skype
  • Viber
  • Workplace from Facebook
  • Hangouts
  • Line
  • Kik
  • Spark

And you can even use Dialogflow to handle one-click telephony:

dialogflow-one-click-telephony

Next to the AI of Dialogflow, this is one of the greatest benefits. You can create a chatbot on almost any channel.

While Dialogflow provides many channels, often you need to use some code to connect your chatbot to a channel. This is is good to know when considering Dialogflow as your chatbot builder.

  • Website
    Can you publish your chatbot on your website?
  • Facebook
    Can you publish your chatbot on Facebook?
  • Instagram
    Can you publish your chatbot on Instagram?
  • Whatsapp
    Can you publish your chatbot on WhatsApp?
  • Telegram
    Can you publish your chatbot on Telegram?
  • SMS
    Can you publish your chatbot on SMS?
  • Slack
    Can you publish your chatbot on Slack?
  • Skype
    Can you publish your chatbot on Skype?
  • Email
    Can you publish your chatbot on email?

Integrations

Unfortunately, Dialogflow doesn’t provide a live chat integration, which is the most important integration of any chatbot builder.

Gladly, there are some companies that provide a combination of live chat and Dialogflow, such as Botcopy.

Moreover, if you want to integrate Dialogflow with other systems, you need to use Fulfillment. With Fulfillment, you can code your own webhooks so you can integrate Dialogflow with other systems.

  • Livechat
    Direct integration with live chat, either via a built-in live chat functionality or a third-party live chat platform.
  • Zapier
    Direct integration with Zapier, one of the largest integration platforms in the world.
  • Make (formerly Integromat)
    Direct integration with Make (formerly Integromat), one of the largest integration platforms in the world.
  • Google Sheets
    Direct integration with Google Sheets, to store or retrieve data from sheets.
  • Webhooks
    Can you call a webhook inside your chatbot flow to send data to other applications?

Marketing

The marketing possibilities of Dialogflow are really limited.

In essence, you can only use Dialogflow Messenger to create a native widget and put it on your website.

With this native widget, you can change the appearance (such as the title and colors) by editing some CSS (mark up language for websites). Also, you can trigger different intents.

  • Facebook website widget
    Embed your chatbot as a Facebook Widget on your website.
  • Native website widget
    Embed your chatbot as a native widget (your own branding & colors) on your website.
  • Landing pages
    Can you create a simple landing page to drive users to your chatbot?
  • Facebook Ads
    Start a conversation with your chatbot when someone clicks on a Facebook Ad
  • One-Time Notification
    Send one-time notifications via Facebook Messenger for a topic that a chatbot user has opted-in for
  • Facebook comments auto-reply
    Automatically reply with your chatbot to users who comment on your Facebook Posts
  • Button creator
    Add a button to your website to drive users to your chatbot
  • Custom URLs for your chatbot
    Generate a simple URL that can be shared anywhere to send users to your chatbot

Pricing

Dialogflow ES is free if you stay under 180 text requests per minute. This means that you can send 324,000 text messages per month and don’t have to pay anything.

If you go above the 180 requests per minute, you need to pay $0.002 per request. Also, if you want audio input or output, you need to pay extra:

dialogflow-es-pricing

If you have a more advanced chatbot and want to use Dialogflow CX, you pay $20 per 100 chat sessions. Also, if you wanto to use audio input or output, you pay $45 per 100 voice sessions:

dialogflow-cx-pricing

In my opinion, Dialogflow is really cheap. For most chatbots, you don’t have to pay anything, you can just use Dialogflow ES for free. Only when your chatbot is really advanced or when you handle a large amount of messages, you need to pay anything.

  • Free trial
    Does the chatbot platform provide a free trial?
  • Free plan
    Does the chatbot platform provide a free plan?
  • Price per month
    What is the lowest price per month for creating a chatbot on this platform?
    0
  • Monthly pricing available?
    Is it possible to pay monthly?
  • Money back guarantee
    Is there a money-back guarantee in case you are not satisfied with the platform?

Analytics

In Dialogflow ES, you have a dashboard that shows the number of sessions and the number of interactions:

dialogflow-analytics

Moreover, Dialogflow has a “Training” feature that shows the messages your chatbot did (not) understand, so you can easily improve the answers of your chatbot:

dialogflow-training-analytics

If you want more advanced analytics, you can use Dialogflow CX, which also provides analytics on Engagement and the different conversation paths users follow:

dialogflow-cx-analytics-path
dialogflow-cx-analytics-engagement

In general, Dialogflow provides enough analytics to improve the answers of your chatbot.

  • Dashboard
    Dashboard that provides general analytics of your chatbot.
  • User input
    Get the input from users that your chatbot didn't understand.
  • Sent messages
    The number of messages that were sent by your chatbot.
  • Open rate
    The percentage of chatbot users that opened a message inside your chatbot.
  • Click rate
    The percentage of chatbot users that click on a button inside your chatbot.
  • Conversion rate
    The percentage of users that converted through your chatbot. For example, the percentage of users that gave their email address
  • Fall back rate
    The percentage of interactions where the chatbot did not know an answer to a question.
  • Custom dashboards/reports
    Can you create your own dashboard or reports?

eCommerce

Dialogflow doesn’t provide any direct eCommerce integrations.

  • Payments
    Collect payments inside your chatbot
  • Shopify
    Direct integration with Shopify to get products, order information and cart data.
  • WooCommerce
    Direct integration with WooCommerce to get products, order information and cart data.
  • Buy directly within the chatbot
    Ability for chatbot users to buy products or services directly from a chatbot
  • Cart abandonment
    Trigger messages in the chatbot when a user abandons their cart on a website or inside the chatbot

Templates & Cloning

Dialogflow provides 45 pre-built agents that you can use to get started with your chatbot.

Also, you can export each agent and share the ZIP with others (which is basically the same as using templates).

Moreover, you can copy your chatbot flows and chatbot elements so you can easily reuse them.

  • Templates
    How many pre-built templates does the chatbot platform provide?
    45
  • Make template
    Can you make templates yourself?
  • Share template
    Can you share self-created templates with others?
  • Duplicate chatbot flows
    Is it possible to duplicate chatbot flows, so you can easily reuse them?
  • Duplicate chatbot elements
    Is it possible to duplicate chatbot elements, so you can easily reuse them?

Optimization

Dialogflow doesn’t provide any optimization, such as A/B Testing.

  • A/B Testing
    Can you optimize your chatbot flows by performing A/B tests?

Support

Dialogflow has extensive documentation on how you can use their platform.

NOTE: I also have some Dialogflow tutorial videos on my Youtube channel, which I definitely recommend you to check out!

Moreover, Dialogflow provides support via email.

  • Tutorials
    Does the chatbot platform provide tutorials on how to use its platform?
  • Documentation
    Does the chatbot platform provide documentation on how to use its platform?
  • Customer support
    Does the chatbot platform provide customer support?

Conclusion

So, should you use Dialogflow to create your chatbots?

Yes, you should.

But in most cases, I wouldn’t recommend using Dialogflow as a stand-alone builder for your chatbots. I would use Dialogflow in combination with another chatbot builder.

Dialogflow is really good at one thing: understanding the user with Google’s AI. But the hard part of Dialogflow is creating chatbot flows, making a personalized conversation and storing user information (you need a developer to do that).

On the other hand, with other chatbot builders (such as ManyChat) it is really easy to create chatbot flows and store user information. So you need to have the best of both worlds. Use another chatbot builder to create your chatbot flows and integrate it with Dialogflow, so you can understand the intent of the user with Google’s AI.

By the way, here is a post on how you can integrate ManyChat with Dialogflow. And here is a video on how you can connect ActiveChat to Dialogflow.

 

Leave a Reply