A chatbot is one of the oldest envisioned use cases for Artificial Intelligence. Alan Turing suggested that instead of debating whether a machine could think, we should measure how it imitates human language. Microsoft’s Xiaoice already passes the Turing test for 10 minutes.
What about your company? Don’t you believe that some of the queries that your company is responding to on social media, webchat or SMS could be easily and quickly responded by chat bots improving customer satisfaction? If yes, do you already have a customer service chat bot or are you looking to build a new one? If you need to find a chat bot vendor to work with and have just 1 minute to solve this, just fill out 4 fields and we will recommend you the most suitable vendors for you.
If you have more time, strap in, because there’s a few too many bot companies out there.
Decide whether to buy or build and cut your search time by half
First of all, you may be thinking what the fuss is about chat bots and why you need to spend time choosing a good chatbot partner. Natural languages seem relatively simple to us humans however they are one of the most difficult areas to master for machines. If you are wondering why, take a look at why chatbots fail. Keeping your expectations low is a good first step.
Though there are many bot companies out there, we can split them in 3 groups:
- Self-service solutions with GUI: Build your bot with using a simple Graphical User Interface (GUI) by dragging and dropping components. This is not the best approach to build a sophisticated bot but you can build a relatively simple bot very quickly.
- Self-service solutions with an API, SDK or library : These solution providers built an API, SDK or a library to provide a framework for your bot to use Machine Learning (ML) and Natural Language Processing (NLP) capabilities. For example, all bots need to understand the intent in user queries. Building blocks for understanding intent such as parsing the user query can be provided by the API. Using the available tools, even novice developers can build bots in a couple hours. However a best-in-class bot would take significantly longer to build depending on what you want the bot to achieve.
- End-to-end solutions: This is how enterprises (and sometimes smaller companies) roll. Specify what your bot needs to do and get a vendor to build it end-to-end.
The critical decision is: What’s the right type of company for you? 3 things will be important:
- What are the required capabilities of your bot?
- Do you have a sizable budget?
- Are you technical?
Once you decide those, the solution is clear in some cases:
- Go for a self service with GUI if all of these are true:
- You have little to no chatbot budget
- You don’t have limited expectations from your bot
- You don’t have/can’t access much technical skills
- Go for a self service with an API, SDK or library if all of these are true:
- You have little to no chatbot budget
- You don’t have high expectations from your bot
- You have or can access some tech skills
- Definitely go for end-to-end service if you said no to all of the above
- For all other cases, I would keep on exploring both options. For example if you have tech skills, it might be worth building your own chat bot with some support from self-service solutions even if you can allocate a sizable budget for it.
In both cases, having a look at the overall industry landscape would be helpful to decide who to work with.
Understand chatbot SDK/API provider landscape
So you believe you or your team can tackle chatbot programming. You are probably right, chatbot development tools in 2017 are so advanced that you can really build an OK chatbot in a couple hours. Let’s look at your options:
Google, Amazon, Microsoft, IBM, Facebook all want to help you build the next greatest bot. Having won Jeopardy in 2010, IBM Watson is one of the pioneers in the field. However today AI is on the agenda of all tech giants. Their solutions have a supporting community and a small army of developers so you won’t need to worry about developing on a platform that can suddenly be deserted.
On top of already familiar names, relatively new companies are also part of tech giants thanks to recent acquisitions. For example, dialogflow (previously api.ai) is part of Google.
Startups are trying various approaches to innovate in this emerging market. For example, graphical user interfaces and even bots to build bots are offered to facilitate bot development.
If you decided that you need to work with a self-service chatbot solution provider, feel free to explore the self service chat bot providers on our platform. If you believe you may need an end-to-end solution, we have a few more insights for you.
Understand different types of end-to-end chatbot providers
We see 4 distinct classes of chatbot vendors.
Customer service experts
Nuance, founded in 1992, is the oldest and most established company on this list. Though Nuance is quite diversified, they were already established in voice authentication and call center software so they have the business relationships and the data to prepare successful systems
Robotic Process Automation automates mindnumbing back-office work typically handled by outsourcing companies. You can learn more from our extensive article on RPA. Since RPA bots deeply integrate with a company’s systems, they can also help chatbots access data and operational capabilities. Workfusion offers both bots and free (free as in beer) RPA services. Integration of both is powerful yet cheap though chatbots are not offered for free by Workfusion yet.
Pure chatbot players from 2000s
Pure end-to-end chatbot providers
Founded since 2010s, they are developing novel approaches to chatbots. They offer end-to-end service and enable enterprises with limited capabilities to develop chatbots. They may lack experience and business relationships but they are plunging ahead with significant funding. Leading examples: Kore.ai, Personetics Technologies, nanorep, Pypestream, Mindmeld, Semantic Machines
Explore all end-to-end chatbot providers
Now that you know about the industry, it is a good time to examine specific players in the industry.
This list is not comprehensive but includes most of the major companies in the space. For the whole list of companies in the space, visit the chatbots section of appliedAI.com
Given their experience in omni-channel customer engagement, Nuance Communications already understands customer service industry and its chat bot Nina boasts Fortune 500 clients like Coca Cola and ING Netherlands.
Founded in 2002, Next IT is the oldest vendor in our database that focuses solely on chat bots. They claim that no other provider of intelligent interface technology has completed more successful implementations for Fortune 1000 companies.
Founded in 2014, Kore.ai is a relative newcomer into chatbots industry however it is already one of the largest players in the space. Kore.ai Bots Platform aims to provide a complete end-to-end solution with the features enterprises require to scale conversational chatbots.
With offices in Europe, North America, Asia and Australia, creative virtual boasts numerous customers including numerous global financial institutions, retailers and government agencies.
Nano Rep has tens of public references from leading corporations and claims up to 70% reduction in live-agent workload thanks to its chat bot. Just 2 months after we published this post listing Nanorep, they were acquired by LogMEIn which provides authentication and connectivity solutions for those who remotely connecting to networks.
Boasting numerous Fortune 500 references, MindMeld was acquired by Cisco in March/2017 for $125 M. With Cisco’s resources, Mindmeld is one of the vendors with strongest support in the space.
For the latest, complete list, please explore the end-to-end chatbot solution providers on our platform.
Now that you know how you will build your chatbot, it is worth thinking about where to test, analyze and deploy it. We have a detailed article on chatbot ecosystem that can guide you. In short, you can consider the messaging platforms Facebook Messenger, Whatsapp, Kik or Slack for deployment. Especially Facebook Messenger has been one of the dominant platforms for bots, helping Facebook team develop a deep understanding around bots.
Acquisition news added