Transaction or Knowledge Chatbot? Moving Beyond the Siri Syndrome
Now that this technology is here, the very first question we need to ask ourselves is whether we're developing the right kind of chatbots and virtual assistants.
Join the DZone community and get the full member experience.Join For Free
Our ultimate test of chatbot intelligence has become a simple, if not nonsensical, question. This "Siri Syndrome" drives our expectations for virtual assistant experiences, but it doesn't have to.
How many times have you been on the phone to make an appointment or book a reservation, and midway through the conversation asked the person on the other end of the line, "What is zero divided by zero?" I'm guessing probably never.
Yet time and again, we test and judge the intelligence of a chatbot by asking a series of unrelated questions — often at random times. We've all asked our devices an inane question out of the blue like "who let the dogs out?" "when will pigs fly?" or even made the statement, "I see a little silhouette of a man" just to hear what it will say (and if you haven't, you totally should).
We do this because it is not only a source of entertainment (see above), but somewhere deep down we are curious to know if a super knowledgeable and sensitive virtual assistant fueled by AI exists that can enrich or replace a lot of processes served by humans today. My colleague, Hristo Borisov — creator of Progress NativeChat — has coined this the " Siri Syndrome." It's our constant drive for true artificial intelligence and it represents our eagerness to check "are we there yet?" Have we reached that moment in time where computers are trained to think and act as humans?
And we have. Sort of.
The short answer is, the technology exists to create the kinds of AI-powered interactions we once only imagined. However, the real question we should be asking ourselves is how best to use the technology. Once we get beyond the entertainment value of asking our virtual assistants to tell us a joke or if we look good today (again, if you haven't you might want to try — the responses are often amusing) we can start to understand the best way to employ the bots to ensure the best return on our technology investment.
Now that this technology is here, the very first question we need to ask ourselves is whether we're developing the right kind of chatbots and virtual assistants. You can answer this question by understanding not only your goal (what you want the bot to accomplish for you) but also the subtle differences between the two distinct kinds of bots: transactional and knowledge.
Transactional vs Knowledge Chatbots
The difference between a knowledge bot (assistants like Siri, Cortana, Alexa, and Google Assistant fall in this category) and a transactional chatbot is that the latter is optimized to execute a limited amount (four to six) of specialized processes that replace the need to talk to an expert or use more complicated UIs such as mobile apps or websites. The knowledge chatbot, on the other hand, supports thousands of processes and in some cases is able to make decisions for you.
Transactional chatbots are trained on top of structured data and can do a set of limited operations. Think of what a bank operator can do for you over the phone: verify your identity, block your stolen credit card, give you the working hours of nearby branches and confirm an outgoing transfer.
On the other hand, knowledge bots are helping you both make a decision and execute it. To be able to "make a decision," the chatbot is usually trained with a vast amount of unstructured and structured data and is trying to produce a response as an expert. To be effective, a knowledge bot can execute hundreds of processes for you. Think of how many things Siri, Cortana, Alexa, or your Google Assistant can do for you. They can schedule appointments, get directions, provide additional information or resources and so much more.
Transactional Chatbots — Simple, Not Stupid
All that said, it needs to be stated that just because a "transactional" bot does not have an intelligence moniker does not mean it can't be smart. Quite the contrary.
A transactional chatbot done right is the epitome of sophisticated simplicity.
Most people think of transactional bots like the chatbots of the early days — the ones that behave like an answering machine. Users will have to follow a predefined path and often the conversation derails very quickly.
That said, you can develop an intelligent transactional chatbot that is able to understand natural language and help a customer perform a set of operations. You can do this by choosing the right natural language processing tools and implementing a professional conversational UI. These two elements combined allow you to direct the conversation flow and positively impact the experience the user has with your bot. The exchange will migrate from one that might devolve into the end-user walking away frustrated, to a productive interaction and a positive user experience.
As an example, if you have a banking chatbot that can do five operations, each of which requires between three and five steps to complete, you get a set of 20 interactions that your chatbot needs to support, and this doesn't include edge cases where user input is invalid or the user switches his intent midway through the conversation. Microsoft, Amazon, and Facebook already provide state-of-the-art Natural Language Processing (NLP) developer tools to help your bot understand user intents. And Progress provides the industry's first conversational UI components and controls with features like built-in suggested actions, hero cards, and more.
How to Choose
So, how do you decide if you need a knowledge bot or a transactional bot? You need to start by asking yourself why you are doing it in the first place and what you truly need to accomplish by implementing a chatbot in your application. Our general rule of thumb is to try to find a process you can improve 10x. That may mean instead of replacing humans completely with AI-fueled chatbots, you simply optimize the workload of the frontline employees by 70-80% with transactional bots.
This approach is already delivering a competitive advantage to many companies in the financial services, healthcare, travel and insurance sectors. These industries tend to have structured processes suitable for transactional chatbots that are currently performed by full-time employees.
Google has clearly done this with Google Duplex. This is truly a transactional chatbot focused and knowledgeable only for a single task — making an appointment over the phone. Google understands that the path to making AI work for us isn't only in knowledge bots. It is also a path that builds task-driven sophisticated bots for a handful of focused functions. They understand the purpose for which each will be used and have created the technology to solve specific problems.
So, look honestly at what you need/want to accomplish. Understand that the drive for true artificial intelligence does not always have to be the end-game — don't get sucked into the Siri Syndrome — and begin to imagine what you as a developer can create as a way to improve a given customer experience 10x using NLP, conversational UI, and your existing skills.
And while you're doing that, make sure you take a break to ask your virtual assistant to tell you a joke. You'll be glad you did.
Published at DZone with permission of Sara Faatz, DZone MVB. See the original article here.
Opinions expressed by DZone contributors are their own.
RBAC With API Gateway and Open Policy Agent (OPA)
File Upload Security and Malware Protection
Getting Started With the YugabyteDB Managed REST API
Top 10 Pillars of Zero Trust Networks