25 Chatbot Platforms: A Comparative Table

by Olga Davydova

Many experts called 2016 “the year of the chatbots.” Thousands of chatbots already help businesses improve customer service, sell more, and increase earnings. This paper reports a Data Monsters overview of research on the best-known platforms for building chatbots.

The relevance of this research is proved by the massive deployment of chatbots. Indeed, today chatbots are used to solve a number of business tasks across many industries like E-Commerce, Insurance, Banking, Healthcare, Finance, Legal, Telecom, Logistics, Retail, Auto, Leisure, Travel, Sports, Entertainment, Media and many others. Gartner Summits [1] projects that more than 85% of customer interactions will be managed without a human by 2020. Chatbots are expected to be the number one consumer application of AI over the next five years according to TechEmergence [2].

Results of the research

25 of the best-known platforms for building chatbots, such as IBM Watson, Microsoft Bot Framework, LUIS, Wit.ai, Api.ai, Chatfuel, and others were studied, and a comparative table was composed.

Link to the chatbots’ comparative table (Google spreadsheet).

Below is an overview of the most popular bot platforms.

IBM Watson

According to the research study by Mindbowser in association with Chatbots Journal [3] IBM Watson is the first choice as a bot-building platform for 61% of businesses. One of the Watson’s most important parts is a Conversation Service. It is built on a neural network (one billion Wikipedia words), understands intents, interprets entities and dialogs, supports English and Japanese languages, and provides developer tools like Node SDK (Software Development Kit), Java SDK, Python SDK, iOS SDK and Unity SDK. IBM offers free, standard, and premium plans [4].

Royal Bank of Scotland launched Luvo, a chat bot taught to answer customers’ questions in near-real time. It routes overly complex requests to advisers using IBM Watson Conversation [5]. Typical questions it answers with staff via the web chat include: I have lost my card — what steps do I need to take now? I have locked my PIN — how do I unlock it? How do I order a card reader for my business?

Microsoft Bot Framework

Forty-one percent of the businesses in Mindbowser’s study [3] said they prefer Microsoft Bot Framework. It has its own Bot Builder SDK that includes .NET SDK and Node.js SDK. The entire system consists of three parts: Bot Connector, Developer Portal, and Bot Directory. The framework provides the Direct Line REST API, which can be used to host a bot in an application or website [6]. It is open source and available to all on Github, and it supports automatic translation to more than 30 languages. Microsoft Bot Framework understands users’ intents. It is possible to incorporate LUIS for natural language understanding, Cortana for voice, and the Bing APIs for search.

Microsoft Language Understanding Intelligent Service (LUIS) uses intents and entities. All its applications are centered on a domain-specific topic or are content related. Active learning technology is one of LUIS’s features. It is possible to use pre-existing, world-class, pre-built models from Bing and Cortana. Models deployment to an HTTP endpoint is a one-click operation; it returns easy-to-use JavaScript Object Notation (JSON). LUIS offers a set of programmatic REST APIs that can be used by developers to automate the application creation process. In addition, several SDKs are available: C# SDK, Python SDK, Node JS SDK, and Android SDK [7]. Supported languages are English, French, Italian, German, Spanish, Brazilian Portuguese, Japanese, Korean and Chinese. LUIS offers free and standard plans.

Wit.ai

Forty-five percent of the respondents in a study by Mindbowser [3] said they trust Wit.ai more than any other bot-building platform. By April 2017, it was used by over 100,000 developers [8]. Wit.ai allows using entities, intents, contexts, and actions, and it incorporates natural language processing (NLP). There are several clients: Node.js, Python, Ruby, and HTTP API for other platforms. It is available for developers to use with iOS, Android, Windows Phone, Raspberry Pi, Python, C, and Rust; it also has a JavaScript plugin. Wit.ai supports about 50 languages, and it is free.

Api.ai

Another conversational platform for bots, applications, services, and devices is Api.ai. Api.ai matches the query to the most suitable intent based on information contained in the intent (examples, entities used for annotations, contexts, parameters, events) and the agent’s machine learning model. Api.ai transforms the query text into actionable data and returns output data as a JSON response object. There are predefined knowledge packages collected over several years. Available SDKs are Android, iOS, Cordova, HTML, JavaScript, Node.js, .NET, Unity, Xamarin, C++, Python, Ruby, PHP, Epson Moverio, Botkit, and Java. Brazilian Portuguese, Chinese English, Dutch, French, German, Italian, Japanese, Korean, Portuguese, Russian, Spanish, and Ukrainian are the languages Api.ai can support. The platform is free. Google bought Api.ai in September 2016. Now it is possible to integrate your Api.ai agent with Actions on Google that lets you build applications for the Google Assistant, which is the only assistant in Google Home. Google Home enables users to interact with services through voice commands.

Semantic Machines

Semantic Machines created a proprietary conversational AI. The features of the system include a conversation engine, speech synthesis, deep learning, reinforcement learning, speech recognition, semantic intent extraction, and language generation (NLG) technology [9]. Semantic Machines goes beyond understanding commands to understanding conversations. The platform is language independent.

DigitalGenius

DigitalGenius developed Human + AI Customer Service. It’s not exactly a chatbot platform, but rather a deep learning agent efficiency tool that works in any language. AI predicts case metadata and suggests the right answers to your customer service agents. AI learns from every agent interaction. It leverages a deep neural network model, word vectors, statistical operations, and deep learning algorithms [10]. Human + AI Customer Service installs as a layer into existing customer service software like Salesforce, Zendesk, etc.

Chatfuel

More than 360,000 chatbots have been created using Chatfuel, serving more than 17 million users globally [11]. A block is a basic building tool of a bot. It consists of one or more message cards that are sent together to a bot user. The next step is to link blocks with each other using buttons in text cards or in gallery cards. Many plugins were developed: Google search, Bing search, JSON API, RSS Import, Subscribe plugin, Digest, IFTTT, Zapier, user input, and LiveChat. Chatfuel supports about 50 languages, and it is free.

Pypestream

Pypestream’s Smart Messaging Platform uses a patented framework of ‘Pypes’ and ‘Streams’ for natural language processing and keyword parsing. An open and flexible API platform allows custom integrations and development of third party connectors, plugins, and extensions. The platform includes the Smart Messaging Framework, Pypeconnect SDK, Pypemanager, the Pypestream mobile app, as well as API plug-ins and integrations [12]. By April 2016, the company had 500 businesses signed up and using the messaging platform, including Washington Gas and Billboard.

Pandorabots

The Pandorabots API allows you to integrate a bot hosting service and natural language processing engine into your own application. Developed SDKs are Java, Node.js, Python, Ruby, PHP, and Go. Pandorabots uses AIML (Artificial Intelligence Markup Language) and includes A.L.I.C.E. (The Artificial Linguistic Internet Computer Entity) — a natural language processing chatterbot. It is multilingual [13]. Common use cases include advertising, virtual assistance, e-learning, entertainment, and education. Academics and universities use the platform for teaching and research. Pandorabots is an AIaaS platform — Artificial Intelligence as a Service.

AgentBot

AgentBot was specially developed for Latin America so it supports English, Spanish, and Portuguese languages. It uses Aivo’s own natural language processing technology. The platform understands natural language, has memory to maintain coherence during long conversations, gathers customer information to deliver customized solutions, applies continuous evolution, and clarifies intents [14]. AgentBot integrates with any CRM, internal system, human chat, or third party application. For integration, REST API is used.

ChatterBot

ChatterBot is a Python library that makes it possible to generate responses based on collections of known conversations [15]. ChatterBot is language independent. The software license is the 3-Clause BSD License.

Chatscript

ChatScript is the next generation chatbot engine that has won the Loebner’s 4 times and is the basis for natural language company for a variety of tech startups. ChatScript is a rule-based engine. Rules are created in program scripts through a process called dialog flow scripting. These use a scripting metalanguage (a script) as their source code [16]. ChatScript engine has many features like powerful pattern matching aimed at detecting meaning, simple rule layout combined with C-style general scripting, built-in WordNet dictionary, extensive extensible ontology, local machine control via popen/tcpopen/jsonopen, structured JSON data reading from websites, and others. ChatScript runs on Windows, Linux, Mac, iOS, or Android. It has integrated tools to support maintaining and testing large systems. UTF8 support allows scripts written in any language. ChatScript works under the MIT License.

Newer startups

Platforms like IBM Watson, Microsoft Bot Framework, Api.ai, ChatScript and Pandorabots were developed ten or more years ago. Therefore, their experience provides the most advanced tools and offers the most flexible solutions for businesses. These platforms make it possible to use different programming languages. Each platform has developed its own SDKs, uses cutting edge data processing and analysis technologies, supports dozens of natural languages, and are already embedded in customer services, sales, marketing, order processing, social media, payment, recruitment and other industries. Just the same, many startups were created over last few years. Some of them grow fast and are already well-known:

Twyla

Twyla learns from agent/customer live chats, blends machine learning and rule-based methods, answers questions, and deflects tickets [17].

Msg.ai

Msg.ai integrates with popular customer support offerings while leveraging intent models and tone classifications [18]. It supports deep learning, interactive smart cards, and A/B testing.

Rasa NLU

Rasa NLU has HTTP API and Python support, intent classification, and entity extraction [19]; it is an open source tool that runs locally.

Reply.ai

Reply.ai is a visual bot builder that easily leverages natural language processing (NLP) engines wit.ai and api.ai for your advanced use cases [20] with analytics about every single response from your users.

ManyChat

The basic functionality of ManyChat lets you welcome new users, send them content, schedule posts, set up keyword auto-responses (text, pictures, menus), automatically broadcast your RSS feed, and much more [21]. No coding is required. It is free.

KITT.AI

KITT.AI built its proprietary ChatFlow platform that enables users to create conversational agents, or smart bots, using a simple drag-and-drop interface that visually describes a dialogue and at the same time implements the flow that can be executed on the server as the dialogue is designed. The features of the platform are hotword detection (no internet required), semantic parsing, natural language understanding, conversational engine (multi-turn support), and neural network powered machine learning model [22]. ChatFlow supports Alexa, Facebook Messenger, Kik, Skype, Slack, Telegram, and Twilio. The platform is free now and lets you sign up for the beta. In the future the team is going to provide free or cheap access for individual developers and small teams, and higher charges for enterprise customers.

It’s Alive

It’s Alive is a free Facebook page chatbot building platform. The core feature of the platform is recipes. This feature enables you to automatically respond when your users write specific keywords or phrases. If your chatbot has missed a keyword, it will be added to new or existing recipes. The platform enables you to send periodic content (RSS) each day, week or month, or to manually send a one-time message to each chatbot subscriber. The team believes in decision trees and buttons that drive users towards the answer they are looking for [23].

Conclusion

If it is necessary for your business to create a chatbot, hundreds of various chatbot-building platforms are available. There are tools that are centered on a domain-specific topic or universal tools, platforms that allow you to program your chatbot using SDKs, platforms that do not require coding, and environments that support English or your primary language. There are paid and free platforms.

Sources

  1. https://www.gartner.com/imagesrv/summits/docs/na/customer-360/C360_2011_brochure_FINAL.pdf
  2. https://www.slideshare.net/TechEmergence/ai-founders-and-executives-predict-5year-trends-on-consumer-tech
  3. http://mindbowser.com/chatbot-market-survey-2017/
  4. https://www.ibm.com/watson/developercloud/conversation.html
  5. https://www.americanbanker.com/news/royal-bank-of-scotland-to-launch-chat-bot-with-ibm-watson
  6. https://docs.botframework.com/en-us/
  7. https://www.microsoft.com/cognitive-services/en-us/language-understanding-intelligent-service-luis
  8. https://wit.ai/
  9. http://www.semanticmachines.com/
  10. https://www.digitalgenius.com/
  11. https://chatfuel.com/
  12. https://www.pypestream.com/platform/
  13. https://www.pandorabots.com/
  14. http://agentbot.net/en/solution/
  15. http://chatterbot.readthedocs.io/en/stable/
  16. https://github.com/bwilcox-1234/ChatScript
  17. https://www.twylahelps.com/
  18. http://msg.ai/index.html#features
  19. https://rasa.ai/
  20. https://www.reply.ai/
  21. https://manychat.com/
  22. http://kitt.ai/#technologies
  23. https://itsalive.io/en/