Chatbots Comparison

Chatbots Comparison
Source - Google Images

In recent years our dependence on artificial intelligence has gone up due to the value that it brings to the table. With improvements in AI, businesses can optimize their resources in a better way than in the past. Chatbots have helped companies decrease the costs that they incur on customer outreach. This article will discuss three chatbot platforms: Rasa open source, Amazon Lex and Google Dialog Flow and their advantages and disadvantages.


The documentation for RASA is available at this link: Installation (

RASA documentation is unambiguous and concise. It provides an interactive python code at this link: Rasa Playground, where developers can play around and learn how RASA functions. It was pretty thorough about how a developer can fully utilize the rasa platform. RASA also provides an easy-to-understand learner's guide to help make RASA chatbots at this link: RASA Learning Center. Here, developers learn about the various components of a RASA chatbot with the help of short videos and small code snippets. RASA also provides a certification exam to provide a credential that the developer understands the Rasa framework and best practices.

To overcome the problem of lengthy documentation, Amazon Lex provides videos that make complicated things much easier to understand and deliver a user-friendly experience that helps understand how developers can build a chatbot using Amazon Lex. Amazon Lex documentation includes advanced features like Lambda functions and Deep learning technologies. First, configure with an AWS account, which makes configuration with Amazon Lex easier and quicker. If you want to check out some code examples, visit Github.

The documentation for Google Dialogflow is available at this link: Google Dialog Flow. The documentation for Dialog Flow has details related to the editions and different virtual agent services and the features they provide. The documentation offers insight into concepts like intents, entities and dialogue control. Also, look at these Dialog Flow tutorial videos to get started with chatbots.


The developer must have a basic understanding of python to make a chatbot in Rasa open source and write it in a code editor as a graphical user interface is not present. RASA has a classifier called the DIET Classifier, a state-of-the-art architecture that outperforms fine-tuning BERT and is 6X faster to train. It runs on the terminal/command prompt on execution.

Amazon Lex has a graphical user interface that makes it pretty user-friendly, which developers can quickly implement and integrate with technologies like meta(Facebook, WhatsApp). Amazon Lex uses automatic speech recognition (ASR) and natural language understanding (NLU), lending developers a helping hand to create a user-friendly and easily implementable chatbot. Amazon Lex consists of inbuilt bots that help modify according to convenience. Amazon Lex uses features like intents, utterances and slots.

Implementing a chatbot in Dialog Flow was comparatively more straightforward, and the graphical user interface simplifies it. Even a person without a tech-related background can easily navigate through the controls. If you are trying to check how it works, you can sign up here and look at the platform's features.

Also, if you want to know more about how to build chatbots for your website without being required to code using Dialog Flow, look at the playlist by clicking here.

Customization and Configuration

RASA is a bit difficult to configure on the fly; however, it provides much customization in terms of business logic. Rasa even provides its SDK to implement custom logic. The developer can specify which pipelines they want to work with, make custom forms and actions and define custom entities. Once the developer gets the hang of it, customizing the chatbot is an easy experience.

As we know, no platform can fulfill all wishes of human beings, but somehow to do customization, features are required to qualify for projects. Other platform technologies give freedom to configure NLU, Core Integration and Deployment. But on the other hand, Amazon Lex doesn't allow any customization in their code, only customize in fulfillment.

In Google Dialog Flow, creating new user intents and entities is easy. We can make various custom intents and train them by providing training data. Also, when it comes to entities, there are custom entities and system entities. For custom entities, we will have to train the model with the words that come under that entity and their synonyms if required. Along with the intents created, we can have follow-up intents that help follow up on user requests.


Rasa open source is free to use. An enterprise edition provides additional features like Analytics and Expert Support. You can find more details at Pricing: Open source & paid enterprise subscriptions | Rasa

Amazon Lex, it's paid for what you use. The pricing is based on the requested quantity. Visit to know more about pricing.

Dialog Flow is priced monthly based on the edition(i.e. CX or ES) and the number of requests made during the month. Visit the following to know more about this: Dialog flow Pricing.


In Rasa, developers can integrate any external API, like SQL, or the developer can integrate their API. Further, Rasa developers can integrate RASA with messenger services such as Telegram and Messenger.

In Amazon lex, you can directly integrate with three social media platforms. They are Facebook messenger, Twillio and Slack. If you want to integrate with other platforms, you will have to use the Lambda function.

Developers can integrate Google Dialog Flow chatbots with voice assistants, text-based apps and other open-source platforms. Integration with platforms like Telegram and messenger was relatively more uncomplicated than other chatbots. But for open-source platforms and a few other platforms, it is required to have a google cloud account to integrate your chatbot into those platforms. Google cloud services will provide a credit for $300; after exhaustion, payment must be made to use their services. Also, in case you want to have a look at how to integrate with Telegram, please visit the following article:

It's tough to decide the best platform for chatbot development as Dialogflow takes more time and resources to be trained but is easy to integrate, whereas Rasa gives more customization functions. Our research concludes that every chatbot is precisely the best in some aspects, but it depends on the user. Each platform has perks, and it's tough to decide on an overall winner. Rather than looking for the best, deciding which platform to use would be better based on your requirements.


P Anuraag Reddy Sumit Dhakad Sunjeevan Nelluri