AI for the Enterprise

3 types of business chatbots you can build

Share this post:

Key Points:
From a business perspective, here are the 3 most common chatbots that are being built:
– Support chatbots that are built to master a single domain
– Skills chatbots that are single-turn type bots that do not require a lot of contextual awareness
– Assistant chatbots that are the middle ground between a support and skills chatbot, knowing a little bit about a variety of topics

Build your own IBM Watson chatbot

 

A few years ago when chatbots were just gaining popularity, there was a lot of talk around what a chatbot actually was. With the advent of natural language processing and various machine learning techniques, some of the more advanced conversational applications wanted to separate themselves from their competition. Many began calling themselves “virtual assistants.” This implied that they were somehow bigger or more powerful than existing chatbots, or perhaps were more conversational or could cover a wider range of topics.

However, we quickly discovered that the market did not care how powerful the bot was or about the underlying technology, so long as it solved the right problems. So in a way, many of these different terms for bots became more or less synonymous with each other. It didn’t matter what you called it – you were getting something you could hold a conversation with. We’re now at a point where we know that regardless of what you call the bot, there are usage patterns and differentiation that make chatbots distinct.

When you’ve done your research and are at the point of beginning to build your bot, think carefully about what problems you’re trying to solve and what functionalities you will want to incorporate. Knowing what you want your application to solve for and assist with will decide the type of chatbot, virtual assistant or agent you ought to build. This will impact both your development plan and, as importantly, your end-user experience. The following are the three main types of chatbots I have come across, with background on their particular uses and variations.

1. Support chatbots

Support chatbots are built to master a single domain, like knowledge about a company. Support chatbots need to have personality, multi-turn capability, and context awareness. They should be able to walk a user through any major business processes, and answer a wide range of FAQ-type questions. You will want to have a short-tail and long-tail combo solution when building this type of chatbot. At IBM Watson, we would use the Watson Conversation service for the short-tail, common questions and processes, and Watson Discovery service for the long-tail, but there are many potential solutions for this. Speech is an optional feature, and not a necessity, since users typically have sat down at a desktop and are ready to figure out their solution. The chatbot developer will want to spend the most time making sure it is as easy as possible to navigate the bot, and ensuring it can execute the actions that your users actually care about (for example, just because you want to sell more credit cards doesn’t mean your customers want to open more credit card accounts).

2. Skills chatbots

Skills chatbots are typically more single-turn-type bots that do not require a lot of contextual awareness. They have set commands that are intended to make life easier: “Turn on my living room lights,” for example. Speech functionality is recommended for this type of chatbot so the user does not need to turn on a device or click any buttons. They should be able to follow commands quickly, so that your users can multitask while engaging with the bot. These chatbots do not need to worry too much about contextual awareness, unless you want to design a particularly advanced one, as people will quickly learn what to say, and say it appropriately. It’s a nice bonus if you can give a command, and your bot knows – to return to our example – that you are in the kitchen and acts to turn on the correct lights.

However, this is not a necessary function, as users will quickly learn to give the appropriately specific command. When building a skills bot, it is important to focus on integration, especially when controlling a home or personalized objects. Keep integration simple so your users can interact with the bot without worrying about how to use .

3. Assistant chatbots

Assistant chatbots are more or less a middle ground between the two bots above. They work best when they know a little bit about a variety of topics. Many people envision these bots will someday become navigators of all other bots that are out there now. Want to pay a bill? Ask your assistant bot to talk to the support bot for your bank. Assistant chatbots need to be conversational and respond to just about anything, while being as entertaining as possible. Siri is a good, current example – while she only does so much, people continually ask her for things simply because even when she cannot perform the command, the response she gives tends to be amusing. When building an assistant chatbot, it is important to make it as obvious as possible how the bot is trained. The range of questions a user might ask is large, so making sure you have adequate coverage is going to be the most difficult factor. In many cases, when people do not know what they should ask, they will not ask anything at all. And if you miss the few topics they initially are willing to try, they will not come back for more.

Even though these are the most common types, many bots in production fall somewhere in between two. Some are even a combination of all three. No matter what type of bot you decide to build, it is important to give your bot some life and personality, make it useful, and make sure it’s easy to use. People interact with bots because they want to get something done in a more natural way than was previously possible. Whether it’s something simple like turning on a light, or something complex like applying for a mortgage, every pattern has specific features that make it stand out, so be sure your bot shines brightly in what it’s designed to do. The possibilities are endless.

For additional bot-building inspiration, check out IBM Watson’s chatbot landing page.

Watson Conversation Offering Manager

More stories
July 16, 2018

Building trust in the age of AI – How businesses can build fairness into their machine learning models

Every organization that develops or uses AI, or hosts or processes data, must do so in ways that allow them to rationalize the decisions or recommendations in a way that is easily consumable. Let's examine Forrester's recommendations how organizations can leverage AI for the good of humankind, while avoiding the ethical pitfalls associated with perceived discrimination.

Continue reading

July 12, 2018

Improving productivity: IBM ranked leader in AI-based text analytics platforms by Forrester Wave

IBM has been named a leader in AI-Based Text Analytics Platforms by The Forrester Wave ™, Forrester's rigorous evaluation of vendors in a software, hardware, or services market. Watson Discovery, Watson Explorer, and Watson Natural Language Understanding’s strengths and weaknesses were assessed alongside seven other providers.

Continue reading

June 26, 2018

Updates to Watson Visual Recognition – 
Price reduction for Custom Classification, and 
Food and Explicit Models are now GA

Announcing updates to the IBM Watson Visual Recognition: A price reduction for Custom Classification events, and two models becoming generally available. Our blog highlights all these exciting changes.

Continue reading