CAI Businesses and Products
Gartner has said that there are more than 2,000 enterprise chatbot vendors. Here I list some of the existing businesses and products that fill the CAI ecosystem and that may be useful for other companies or researchers who want to learn about or implement conversational artificial intelligence (CAI). I do not claim to list all of them. I will start with what I know and continually expand this list but focus on what I believe are the most unique or the most important in their areas of the ecosystem.
Some industry analyst categorize CAI-related businesses into chatbots, messaging apps, digital/personal assistants, and voice search. I further divide the market into the following.
- The Best Chatbots
- Custom Chatbot Development
- General CAI Toolkits, Frameworks, and APIs
- Customer Support Agents
- Smart Speakers
- Voice Assistants
- Messaging Applications
- AI Avatar Studios
- Speech-To-Text Systems
- Text-To-Speech Systems
- Industry Research, News, Education, and Publication
The Best Chatbots
The purpose of CAI is to create chatbots. Most chatbots are accessed on what I call chatbot platforms such as virtual assistants or messaging applications (see below), or more directly on individual organizations’ apps and websites.
There are already so many out there that I will not attempt to list them all. However, I will try to list the most notable ones.
Mitsuku (by Pandorabots)
Mitsuku, a record breaking five-time winner of the Loebner Prize Turing Test, is considered by some to be the world’s best conversational chatbot.
Custom Chatbot Development
Garnter advise businesses to not attempt to create chatbots unless you have the right skills. If you do not have the skills, there are plenty of third-party development and hosting companies that you can pay to do the work.
RAIN’s mission is to help brands innovate at the intersection of marketing and technology. They define voice strategies, design conversational experiences, and build voice technology software. Their team is equal parts creative, strategy, and engineering, with offices in New York City, Utah, and Seattle. They work with the world’s most ambitious brands across finance, consumer products, hospitality, healthcare, entertainment and more.
RAIN has closed a $3 million Series A investment led by Stanley Ventures to build on this foundation to enable RAIN to create conversational AI solutions for the “deskless workforce” — whose labor powers industries like agriculture, healthcare, retail, and manufacturing — with the goals of streamlining laborious processes and increasing enterprise productivity.
“Founded in 2017 and headquartered in San Francisco, CA, SmallTalk Agency specializes in customer engagement and experience. The company builds enterprise chatbot solutions on a variety of messaging platforms utilizing AI, machine learning, and NLP, and it can integrate with a variety of CRM systems and social media platforms. Recently, the company designed a Facebook Messenger-based chatbot for startup Next Insurance through which customers can shop for coverage.”
“Based out of Dallas, TX and founded in 2008, [x]cube LABS is a mobile engagement and development company. [x]cube LABS offers, among other digital applications, chatbots that leverage machine learning and speech recognition. The company cites insurance industry chatbot expertise, and clients include UnitedHealthcare. Specific insurance use cases include claims management, mobile policyholder self-service, and underwriting.”
General CAI Toolkits, Frameworks, and APIs
Alexa and Lex (by Amazon and AWS)
Avaamo CAI Reference Architecture
Bot Factory (by Emotibot)
Founded in 2015, focused on Chinese market (as of 2019), Emotibot makes a CAI development platform targeting large and medium sized companies across industries. They emphasize emotional artificial intelligence and the ability to customize without needing to code. They generalize across text, voice, and vision input.
- Real-time knowledge recommendation for CSR
- Phone Bot for automatic outbound calls
- Multi-Turn Contextual Dialog Flow Management
- Intent Engine
- Knowledge and Inference Engine
- Emotion Engine
In the news:
- https://www.prnewswire.com/news-releases/emotibot-technologies-raised-45-million-in-a-series-b-round-launched-ai-contact-center-with-full-duplex-voice-conversational-ai-300946802.html (This one was hard to read through the Chinese translation and marketing hype)
Cognigy provides an AI-based enterprise platform to build, deploy and maintain conversational automation solutions. Its technology connects operational touchpoints through out-of-the-box integration to business intelligence (BI), customer relationship management (CRM)and enterprise resource planning (ERP) tools. Users can access real-time data through simple conversation. From customer service questions to high-level IT operations management, Cognigy enables a fast and effective way to deliver results through the most natural channel — conversation. Cognigy focuses on CAI builders who are not engineers such as conversation designers and project managers. It can be deployed on premises or in the cloud and for both customer- and employee-targeted self-service support. Businesses maintain centralized control of security, user administration, privacy and compliance.
DialogFlow (by Google)
One of the oldest and most mature. Might also be more costly.
Speech-to-meaning, knowledge graphs, context, crowd-sourced domain knowledge.
- Website: https://www.houndify.com/
Kore.ai is the industry’s first and only enterprise-grade, end-to-end conversational AI-powered bots’ platform to design, create, train, test, and host AI and NLP-powered chatbots for use in the most popular consumer and business communication channels.
Kore.ai is primarily offering a comprehensive, end-to-end platform (hosted as well as a service) that allows enterprises to build and deploy out-ofthe-
box or completely customized chatbots for their customers and workforce.
Microsoft Bot Framework / LUIS (by Microsoft)
Bot Framework for deploying bots. LUIS for NLU
MindMeld (by Cisco)
Founded in 2011. Python-based open-source chatbot framework. Comparable to Rasa — more mature but has less of a community.
Nuance Communications, Inc.
Nuance earned a differentiated rating from Forrester (2019) in the artificial intelligence (AI), voice and speech, human/AI blending, omni-channel and security and authentication criteria. Nuance is distinctive in its interweaving of humans and AI compared to rivals with its tools for human agents. Nuance did not receive any “Needs Improvement” scores in the evaluation. Nuance has focused on improving NLU, biometrics, conversational design, and agent AI. They also seem to focus on global financial and telecommunications companies.
Stats. Nuance serves over 30 billion customer interactions per year. Over 400 million consumers make more than 8 billion successful authentications yearly using Nuance biometrics. Nuance Conversational AI solutions have delivered organizations over $3 billion in annual savings.
OpenDial spoken dialog system
Academic, Open source
Intent recognition that is pre-trained and then fine-tuned on your own data.
Open source toolkit for statistical dialog systems. Contains several RL-based dialog policy learning algorithms. Contains a benchmark environment for dialog evaluation. From Cambridge University and Milica Gaši
Founded in 2016. Written in Python. Open Source CAI framework. Great community and developer education resources.
Read more in TP on Rasa.
Simple Chat UI in Python
Teneo by Artificial Solutions
Conversational AI platform. Focus on large B2C enterprises which build multiple conversational assistants. Graphical and collaborative UI. Predicts that there will be a tipping point transitioning CAI from an early adopter market to an essential tool for all companies to interface with their
customers. They claim they have built systems that perform better than humans: no boredom, no inconsistencies. No employee turnover. Reduces cost and provides more data about their customers to learn from as a brand.
Interview with CEO of Artificial Solutions about IPO and business: https://www.fool.com/investing/2019/02/25/a-conversational-ai-leader-is-goingpublic.aspx
Telegram Bot API
Web login: https://web.telegram.org/#/login
API Docs: https://core.telegram.org/bots/api
Python how-to: https://djangostars.com/blog/how-to-create-and-deploy-a-telegram-bot/
Customer Support Agents
IBM Watson Assistant
Based on https://www.passage.ai
Oracle Virtual Assistant
Stephen Stroup’s groups is considering using this as it is part of their Customer Management System
Other Notes on Ecosystem
Provider of AI-powered virtual agents for customer service. SmartAction uses proprietary conversational AI to automate conversations that used to be handled by call center agents over phone, chat, and text. Every SmartAction solution bundles robust end-to-end customer experience (CX) services with their SaaS offering.
These are specialized voice assistant hardware.
Also called digital assistants, virtual assistants, or personal assistants. These are voice-enabled applications that run on general hardware like smartphones as opposed to specialized hardware a.k.a. smart speakers. They are easy to use by voice but they have challenges with authentication and push-style interaction.
Google Home Assistant
Instant messaging applications are one platform on which to develop chatbots. They are good at both push and pull style interaction.
This Google product is going away in the future.
AI Avatar Studios
ObEN PAI Studio
The world’s first AI avatar creation and management platform. Personal AI (PAI) Studio will enable brands and creators to design, build, and manage avatar-based, AI content via a turn-key software solution.
Google Cloud Speech-To-Text
Microsoft Azure Speech-To-Text
Mozilla DeepSpeech + Common Voice
Google Cloud Text-To-Speech
Microsoft Azure Text-To-Speech
One of the Open Source Tacotron implementations
Industry Research, News, Education, and Publication
Independent research, online publication, newsletter and podcast focused on the voice and AI industries.