Two common tasks for a person in every business environment are looking up a piece of information and checking the status of a task. How do you usually go about answering these questions? Since it’s the easiest, you probably ask your coworker sitting next to you, or your assistant if you’re lucky enough to have one. Or you might want to stretch your legs, take a stroll to another department, and ask an expert such as your IT staff, AR/AP coordinator, or HR rep. However, let’s face it, most of us would do any of the above to avoid jumping through hoops of logging into a slow legacy system and looking up the information ourselves.
Think about how quickly (or slowly) you might be able to check your utility bill, or make an appointment with your physician – either by using a mobile app or calling your provider on the phone. It’s usually a 5-10 minute task, with a number of potential hurdles – will involve at least a few of these actions:
Now imagine instead, you could simply “ask a friend," your handy Virtual Assistant: “Can you book me an appointment with Dr. Brown on Thursday next week?” or “What was my electric bill last month?” Instantly, you receive your answer or confirmation. The good news is, the capabilities and technologies have arrived to enable these experiences, and they’re poised to change the way businesses and customers interact.
You’ve likely heard some hype around “conversational interfaces” or “chatbots” (for example Apple’s Siri, Amazon’s Alexa, or text-based bots on Facebook Messenger or Skype). These technologies all use a similar interaction model in which users primarily interact with their natural language – either spoken or written. Interactions can range from simple commands and requests to highly complex, multi-branch scenarios. Generally, conversational systems intend to simulate a human conversation, driven by data, that provide personalized, contextual responses to a given task at hand. Many businesses have started taking advantage of these technologies to drive operational efficiency, improve customer experiences, and provide new, innovative customer interaction channels.
There has been quite a bit of confusion as businesses attempt to navigate the conversational technology landscape – keywords such as “chatbots,” “virtual assistants,” and even “artificial intelligence” and “robotic process automation” (RPA) have been conflated over time. Some level setting around key terms may be beneficial to understanding the space.
Conversational technologies can generally be divided into two categories: voice assistants and chatbots. Put simply, voice-based interfaces (e.g. Siri and Alexa) are referred to as “Voice Assistants” and text-based messaging-centric interfaces (e.g. bots that are embedded into Skype, Facebook Messenger, or a web-based chat on a website) are referred to as “Chatbots”. Many other terms tend to act as catch-alls, encompassing both text and voice-based interfaces – such as “Intelligent Digital Assistants”, “Virtual Agents”, “Conversational Assistants”, or similar variations. Some virtual assistant platforms, such as Microsoft’s Cortana are actively blurring the lines between voice and text-based interactions by offering both capabilities within a single conversational interface.
These conversational systems have also often been confused with “Artificial Intelligence” (or AI for short). The culprit? The undeniable influence of movies and television. Think Star Trek: talking to a computer must be AI, right? While these systems do make use of AI technologies for voice recognition and natural language processing, there is actually very little “intelligence” being used by conversational systems at this time. The technologies just aren’t there yet. “Artificial Intelligence” actually refers to the capability of a computer system to perform tasks that normally require human intelligence, often learning from previous experiences and decisions to adapt and refine results over time.
Common use cases for AI include visual recognition, language translation, or decision making – such as fraud detection, sentiment analysis, or providing product recommendations. AI can be used within conversational systems though, for example: If you order a product from Alexa, Amazon might use an AI model to suggest other products you might like to purchase – driven by data on customers who have similar purchase histories. In summary, AI can be used within conversational systems to drive decisions or perform voice recognition, language processing, etc. – but conversational systems are not simply “AI” (at least not yet!).
Robotic Process Automation (or RPA for short) is another term I’ve seen significant confusion around in relation to chatbots – probably due to the use of the word “robot.” The term has little to do with conversational interfaces, and primarily refers to a set of technologies that enable a virtual “robot” to interact with existing web and desktop applications. RPA tools literally automate the process of clicking and typing, for the purposes of performing repeatable tasks usually performed by humans – entering data, performing workflows, and communicating or passing data between systems without using traditional systems integration tools (i.e. web services, APIs, file drops, etc.). RPA works great as a quick way to integrate systems and automate tasks without costly systems integration, but has little to do with conversational interfaces. However, RPA can be used within a conversational system, for example: a user might use a chatbot or voice assistant to book a flight – the chatbot can call out to execute an RPA process on the back-end to click around and enter data within another system to book the flight. Unlikely, but certainly possible.
Why are conversational technologies just starting to finally take off? Vendors are coming out of the woodwork to provide solutions and platforms to build these new experiences, and 80% of businesses are currently using, or planning to use these technologies by 2020. If holiday device sales are any indication – consumers appears to be readily embracing these technologies, happily introducing virtual assistant devices like Amazon’s Echo Dot and Google Home into their living rooms, offices, and bedrooms.
One key reason conversational technologies are taking off lies within recent technology advancements. The basic components of these technologies (voice recognition and chat bots) have been around since the 1990s. But they weren’t always very good. Voice recognition was relatively poor, used primarily in specialized scenarios such as dictation. Voice recognition has improved rapidly over the past several years due to advances in machine learning allowing systems to learn and iterate, improving over time as new data and voice samples are ingested and processed.
Ever wonder why Google entered the voicemail space with Google Voice? The fruits of those labors are baked into today’s voice recognition in Google Home. Natural language processing (NLP) has also improved by leaps and bounds in recent years, introducing new capabilities like concept identification and knowledge graphs to improve systemized understanding of language. These new language analysis capabilities allow chatbots to more accurately understand context within a conversation, attempting to understand a user’s intent rather than searching for keywords and phrases, making them more flexible and easy to use.
Further driving this trend, conversational technologies have started to enhance and even entirely supplant the mobile app experience. Users are increasingly embracing these technologies on their smartphone devices (i.e. Siri, Google Assistant) to perform actions and request information, rather than opening up an app and jumping through the aforementioned barriers. Users are also interacting with a decreasing number of mobile apps, with many using less than 20 apps on a regular basis. With messaging apps often taking several spots within those top apps, businesses are responding by embedding their user experiences into the apps where users are spending most of their time–enabled by the new chatbot capabilities within Facebook Messenger, Skype, and other apps.
Users, especially younger ones, are increasingly preferring to interact with these technologies preferring instant messaging to phone conversations, and preferring to solve customer service issues without human interaction. Conversational technologies can offer an improved experience over traditional mobile interactions by using data and device sensors (e.g. sound, location, vision) to provide a more personalized, contextual, and immersive experience for users. It follows that the users who are currently growing up with these technologies in hand are more apt to use them as time goes on.
While conversational systems have existed in some form for nearly three decades, the technologies are finally straightforward enough for businesses to deploy and less frustrating for end users to interact with. Numerous conversational platforms and Software-as-a-Service offerings are now available, making these systems easier to implement, deploy, scale, manage, and integrate with existing systems. New features and capabilities on these platforms are being released on a weekly basis, continually improving the user experience. Conversational interfaces have already proven disruptive to traditional mobile app-based interfaces of recent years, proving easier to use with less friction in many cases. Businesses must prepare to take advantage of these new systems to drive operational efficiencies and improve customer experiences, just as it was important to embrace mobile apps and websites in the past.
If you would like to learn more about conversational technologies, feel free to reach out to us for more information. We have experience working with clients to plan, design, and build conversational systems.
I am even more accessible than the other modals.