The ’80s had the fight between VHS and Betamax. The ’90s were more about Blu-ray vs. HD and since the new millennium, we have seen a battle between Android and iOS. So, will the next decade see a similar conflict between chatbots, virtual assistants (VAs) and Virtual Human Agents (VHAs)?
On one level, you could say that the answer is ‘yes’. To some degree, chatbots, virtual assistants, and VHAs have overlapping abilities. The borders between them can to industry outsiders at times appear flimsy. However, there are some key technical differences between the three that are much bigger than the ones between, for example, iOS and Android- two systems developed to essentially do the same thing. For example when it comes to how they leverage artificial intelligence (AI) and/or human-like AI.
What are the differences, what do they mean for the future of chatbots, virtual assistants and VHAs, and what do those differences mean for you as an individual, a company or an IT developer, you ask?
We’re glad you did because that’s what this article is about.
The category that most people will likely be most familiar with is chatbots. While the technical sophistication varies, chatbots are generally automated or pre-programmed interfaces. They usually function as part of messaging platforms, for example on social sites and messaging systems, including the likes of Facebook Messenger, Skype, and Slack, or on company/organization websites.
Chatbots’ core strength is that they can help automate basic tasks and interactions with users and customers. Their main weakness is that their technical limitations mean that they aren’t able to deal with natural language processing (NLP). Most chatbots rely on limited, open-source NLP libraries that haven’t been customized to match the needs and conversational styles in a given industry or by a specific set of users/customers.
As a result, chatbot interaction tends to rely on predefined, structured language and follow a decision tree system that is a bit like a mix of the old ‘Choose Your Own Adventure’ books and an automated phone service. In other words, humans need to ‘speak the chatbots’ language’. If you, or your choice of language, stray too far from the norm/expected, a chatbot is often unable to understand what you want it to do/help you with.
The use cases for chatbots include customer support & relations, initial sales inquiries, interactive FAQ, and appointment bookings.
A potential advantage of chatbots is that their technology and software architecture make them relatively quick and easy to build and roll out. This is somewhat counterbalanced by a lack of scalability and limited opportunities for further developing their capabilities.
Virtual assistants (VA) generally speaking leverage advanced natural language processing to make interactions more dynamic and natural than is the case for chatbots. The underlying systems are sometimes referred to as conversational artificial intelligence. While the level of AI can vary from system to system, they are often capable of automating a wide variety of tasks without sacrificing accuracy. Examples include Amazon’s Alexa, Apple’s Siri, Google’s Assistant, and Microsoft’s Cortana. As illustrated by these four, VAs can serve as centralized control hubs that make you able to control other systems.
VAs also tend to be cross-platform systems that you can use while on your phone, via smart speakers and other, similar platforms. VAs are capable of maintaining consistency across those channels. VAs can also integrate with other systems. For example, the Japanese messaging platform LINE’s bot can connect with things like Google Calendar.
Some VAs are general use but others have specific functions, like helping you schedule meetings or handling emails. They are capable of handling every step of a given process from initial contact to resolution.
While VAs vary, they are often bodiless, meaning that you interact with a voice system through a channel such as a smart speaker or similar, but there is no physical (virtual reality or augmented reality, for example) embodiment of the VA. in other words, you can hear it, but you can’t see it. In other words, while virtual assistants use AI, it’s hard to talk about them as human-like AI systems.
The way we interact with VAs is based on dynamic dialogue where the system is able to understand a wider variety of natural language. That being said, VAs generally don’t understand every nuance of human language, and very few VA systems are able to decode other kinds of communication such as a user’s tone of voice or body language.
Virtual Human Agents
Virtual Human Agents (VHAs) are the most advanced form of the three. Like VAs they are powered by AI-systems. These systems tend to be more advanced, making the VHA capable of more nuanced communication, as well as the ability to take on a multitude of different job tasks. VHAs are the most human-like AI systems around today.
Both VAs and VHA have the ability to learn more about users over time, with VHAs being able to make more nuanced conclusions. Via all possible AI approaches, including machine learning, they can improve how well they perform tasks over time. In essence, they improve every time they interact with a human. The data gathered and used to deliver the best possible can, as is the case for Connectome’s VHAs, include body language, the tone of voice and other factors.
VHAs tend to be ‘embodied’ entities. In the case of Connectome’s VHAs in the form of a human avatar. A user can see the avatar, which can both read and display emotions via facial expressions and body language. The avatar can move with the users, including across devices. This creates a more natural human-technology interaction that increases trust and also makes communication more nuanced.
Use cases are extremely broad. A VHA could work as a receptionist in a company, be a personal assistant that handles many different tasks for an individual, and work in healthcare-related fields, to mention but a few. We have addded a couple of links to articles about use cases for VHAs at the end of this article.
The increased functionality and data gathered calls for extra protection, which is why Connectome uses blockchain technology to both safeguard users’ data and find new ways of improving the VHAs’ functionality.
Thanks to the broad scope, VHAs are also an excellent choice for developers looking to create solutions for companies, individuals or organizations. Collaborating with monetary incentives and gaining worldwide exposure for solutions can be stumbling blocks, which is one of the reasons why we are developing the Connectome Marketplace. Here, developers and creators can gain exposure for their creations, ensure ownership rights and earn capital from their hard work.
Pick And Choose
It is important to note that the exact distinctions between the three categories presented above is debated. However, it is generally accepted that defining differences between them include complexity, adaptability and use cases. Those use cases for individuals, companies and organisations are not mutually exclusive. In other words, you can invest in – and use – all three if you feel that it makes sense to your situation.
While chatbots dominate the market today, their relatively low functionality and scalability limits their growth potential and use cases. VAs have a broader use, and they could dominate developments in the short term.
However, the increasing number of internet-connected devices and the growing need for some form of intermediary between ourselves and the increasing number of electronic systems we interact with leads to the conclusion that the limits to VAs capabilities make them less suited for the future of human-technology interaction.
VHAs, on the other hand, have almost unlimited potential, which is one of the reasons why we at Connectome believe that they, not chatbots or VAs, will be the technology that will come to dominate in the future.