Platform to Choose for NLP Based Chatbot
Providing excellent customer experience is essential in a business in this modern era of technology, but without sufficient manpower, it seems quite difficult. Don’t worry. Chatbots are here to rescue. Chatbots have become quite prominent in the case of user experience. Nowadays, companies are getting more and more dependent on chatbots in terms of automating and reducing the workload on humans. But it might be difficult to choose which NLP-based chatbot platform is the best according to features, channels, pricing, language support, etc. It majorly depends on what type of bot you are asking. To help this process, I will try to give a summary on how to build a chatbot on a whim or with less coding experiences using some of the topmost NLP-based chatbot frameworks like Dialogflow of Google, Watson of IBM, Lex of AWS, Azure bot, RASA, and cognigy. This article will focus on evaluating these platforms based on some criteria rather than building a bot from scratch.
So, without further adieu, let’s get started.
Dialogflow
Dialogflow is a natural language understanding platform that helps to design and integrate a conversational user interface into mobile apps, web applications, devices, bots, interactive voice response systems and related uses. Dialogflow provides a web interface to create bots making it easier for even non-techies to create bots. The basic attributes such as Intents, Entities, Actions provide engaging ways to interact with the users. You could design a basic chatbot with Dialogflow in a matter of hours.
Dialogflow is a cloud-enabled solution provider. Dialogflow provides integration with Google Assistant, websites, Slack, Facebook Messenger, Skype, Twitter, and many others. At present, Dialogflow supports 25+ languages.
Setting up sessions or updating an agent is absolutely free. Standard plans including text or voice recognition start from $0.007 per request or $0.06 per minute. You can go through the pricing here.
Pros:
- Easy to understand and integrate
- Great explanatory documentation
- Real-time conversation testing option
Cons:
- Not much, but sometimes documentation might seem vague.
Watson
Watson is capable of answering questions posed in natural language which is developed in IBM’s DeepQA project. With the help of a knowledge base, Watson provides a platform easy to navigate. Creating a skill, you can deploy Watson in any cloud or local environment.
You can connect Watson with Facebook Messenger, Slack, Voice Agent (Telephony), WordPress plug-in, and also custom applications through APIs. It supports 10+ languages.
The lite version is free and the plus version starts at $140 per month. See pricing here.
Pros:
- Easy to understand and integrate
- Automated predictive analysis and tone analyzer
- Great documentation
Cons:
- Costly compared to Dialogflow or Lex
Lex
Amazon Lex offers a platform to build bots that automates simple tasks and drives operational efficiencies across the enterprise. It comes with an easy-to-use guide that takes you through the steps of creating your own bots. The backend infrastructure of Lex is supported by AWS Lamda which is widely regarded as the most advanced server-less computing platform.
Amazon Lex integration support is limited to Facebook, Kik, Slack, and Twilio SMS. It currently supports six languages.
Voice requests are charged at $0.004/request while text requests are priced at $0.00075/request. Know more about pricing here.
Pros:
- AWS lambda supported
- Automated speech recognition
- Intuitive testing platform
Cons:
- Lesser language support than Dialogflow or Watson
- Web integration is more complex than Watson, critical data mapping
- Needs a little heads up for following documentation
Azure Bot
Azure bot services allow the building of a basic bot to alleviate the high volume of inquiries and control of customer data. It also gives a virtual assistant tailored to one’s expectations or brand name. It also has some existing bot services. The azure bot supports both clouds and on-premise solutions.
The chatbot created through Azure Bot Service can be published to different channels such as Web, Facebook Messenger, Skype and Skype for Business, Microsoft Teams, Slack, etc. It supports more or less 20 languages.
Azure bot service comes with a free plan of 10,000 messages/month. Post that, pricing is $0.50 per 1,000 messages. Apart from this, they also charge you for resources consumed on Azure functions and the Azure web app. See details.
Pros:
- Microsoft Luis supported
- Usage of existing directory
- Pre Built entities
Cons:
- It supports C#, Java, Python, Javascript for the development
Cognigy
Cognigy is an enterprise conversational automation platform that offers a free trial to test, compare and decide if the platform is right for your purposes. It requires almost zero coding skills and can build a bot through visualization. It supports both on-premise and Cognigy SaaS cloud deployment.
Cognigy endpoints can integrate your bot to voice gateway, web chat, Facebook Messenger, Twilio, Slack, Azure bot, Dialogflow, Alexa etc. It supports 23 most common languages.
Pros:
- Code-free development
- Pre-configured bot
- Sentiment analysis and speech recognition
Cons:
- No access control and open API
RASA
Rasa (NLU + Core) are open source python libraries that require slightly lower level development. Rasa NLU handles projects/intents/entities whereas Rasa Core handles dialogue and fulfillment. Rasa does not provide hosting outside their enterprise offerings but can be operated on-premise.
Rasa supports 7+ languages. Pricing varies based on the plans offered.
Pros:
- Most interactive documentation
- Open source and self hosted
- Docker image
Cons:
- Difficult to configure on the fly
- Need some Python coding
A summary of what I have already discussed:
Finally, the big question is, which platform to choose? It depends on how you are going to use your bot. If you want less coding and design a bot on the fly, then Cognigy, Dialogflow could be a better choice. Amazon Lex has the most user-friendly experience. So it will be easy enough if you can narrow down your preferences. This feature comparison will give you a head start. Let me know which chatbot you find interesting to use and configure.
Thank you for reading. Let’s build some chatty machines.
References
- Google Dialogflow — https://cloud.google.com/dialogflow
- Amazon Lex — https://aws.amazon.com/lex/
- IBM Watson — https://www.ibm.com/cloud/watson-assistant
- Azure Bot — https://docs.microsoft.com/en-us/azure/bot-service/?view=azure-bot-service-4.0
- Cognigy.ai — https://docs.cognigy.com/docs
- Rasa — https://rasa.com/docs/2
- Dialogflow vs Lex vs Watson vs Wit vs Azure bot — https://www.kommunicate.io/blog/dialogflow-vs-lex-vs-watson-vs-wit-vs-azure-bot/
