By 2021, 15% of all customer service interactions will be completely handled by AI. This means an increase of 400% from 2017. It’s not us saying that, but Olive Huang, research vice president at Gartner


When we talk about AI customer service interactions, we are mainly talking about chatbots and virtual assistants. Those are nowadays everywhere, although customers’ expectations about those technologies haven’t been exactly met, casting a shadow of doubt over the capability of such technologies to have a significant impact and real use in the future.


Who hasn’t felt exasperated because you tried to reach customer service online just to be faced with a limited-menu chatbot answering? Who hasn’t felt the need to scream “I want to talk with a real person” to a virtual assistant (VA) reading you endless dial options on the phone? Despite those annoyances, during the past two years, the number of organizations which have developed artificial intelligence services has grown by 10%.



Gartner is known for launching its Gartner Hype Cycle every year, in order to predict the stages of adoption for the most critical technologies for human innovation. On the 2019 release of its Hype Cycle for Artificial Intelligence, Gartner shows how chatbots and virtual assistants are “going down” in the hype cycle, moving across the Peak of Inflated Expectations to the Through of Disillusionment. However counterintuitive, this process is desirable, because it means the technology is able to set realistic expectations concerning its development and application, and that it is actually moving towards the Plateau of Productivity.

Hype Cycle for AI 2019

The base technology and methods used in developing bots and VAs is bound to see a fast evolution over the next few years which, in turn, will help adjusting the users’ expectations and set up actually realistic use scenarios. As AI becomes increasingly embedded in chatbots, VAs, conversational UI and the IoT, the occurrence of self-service will soar.

With SuperCards we are also moving one step closer to that reality. SuperCards are not just another bot or VA: compared to a “traditional” bot, SuperCards are actually usable, they support business processes and can be easily tailored for supporting those, and provide users with timely insights to take action.

Chatbot? Virtual assistant? Who is who in AI customer service?

Chatbots and VA are the future of AI customer service interactions and yet, it is incredibly easy to mix up those types of software, their capabilities, and most promising applications. This is because the lines that separate chatbots and virtual assistants are really thin and, sometimes, blurry. Let’s take a closer look to sort this out:


A chatbot is an artificial intelligence (AI) software that can simulate a conversation using natural language in diverse platforms such as messaging apps, websites, mobile apps or telephone calls. 

Modern chatbots are really useful and most frequently used in simple use case interactions where only a limited range of responses are needed.

This currently means they are used for customer service and marketing applications, where the chatbots are expected to be able to provide answers to questions about the products and services of the company or its policies. However, when the questions exceed the abilities of the chatbot, usually the customer is “upgraded” to a human operator. This is why we believe that the  current “intelligent” chatbots on websites will be replaced by more and more human interaction, as it is too difficult to handle all the customer service use cases with just one chatbot logic, as is just too much to handle.

As we have hinted above, the most common use cases for chatbots are usually related to customer service, and not that much for internal company users. However, there’s no good reason why, in business software, we couldn’t use these same technologies to help the employees to be more efficient. In this scenario, the use cases would then be already naturally known as they would align with the business processes and tasks that the employees are doing anyway.

It is true that there are a few chatbots and assistants for business software in use nowadays, such as MS Teams integration for Dynamics 365, but they all have one problem: they are too generic and cannot be modified to company specific needs. In addition, they just update the information to one system, but moving data between different software solutions is not possible and that would anyways require for custom programming.

As a result, companies haven’t yet invested in these chatbot technologies for internal use, since it would often mean to also develop a quite costly integration project. They need a solution that takes care of the integration part as well and they need to be able to utilize Zapier-like instant integrations with the chatbot technologies. And that is our goal here in Supervise.

Virtual assistant

Virtual assistants (VAs) or virtual customer assistants (VCAs) are, like old-fashioned analogic assistants, a software-based digital agent that supports the development of daily tasks, acting on behalf the organization or one of its individuals on the role of engaging, delivering, informing and acting towards the customer, the own organization or a third-party. Therefore, they are used both internally and externally in the organization, unlike chatbots that are mostly used externally.

Virtual assistants can, for instance, take notes for us during a meeting, remind us about due dates, set alarms, make bookings, schedule appointments, read out loud messages sent to us, control and manage smart devices, provide directions, make calls and so forth.

As a result of their rich and multiple capabilities, VCAs require much more infrastructure and a more complex coding logic than chatbots. Besides, they keep a memory of past interactions and can form long lasting relationships with the customers, building a rich user experience that creates business value and engages the customer.

However, currently not all VCAs are succeeding in creating a compelling user experience and delivering business value. Failed implementations that are uncapable to decipher intent or handle uncommon input efficiently are often source of frustration for the customer and won’t survive for long on the current trend that pushes towards a seamless omnichannel engagement.

Chatbot vs virtual assistant in AI customer service

In short, while chatbots are often used externally as information acquisition interfaces that depend on NPL algorithms and which cannot interpret intention or the nuances of natural language, virtual assistants are used both internally and externally for support tasks when conducting business and have the ability to maintain the conversation flow even after the interaction has been broken. While both chatbots and virtual assistants are conversational interfaces, their capabilities and applications differ significantly.

How can I get my own customer service chatbot?

It’s likely that by this point you have been convinced of the necessity of implementing AI customer service technologies in your company. Excellent! You are one step closer to providing a compelling and engaging user experience! However, before dipping your toes into the rough waters of AI technologies, we recommend you to take a big breath and consider the following:

– What is the goal of using this technology?

Are you looking for a customer service tool? Or do you also seek support on internal tasks? A clear definition of the purpose will help you determine whether you need a chatbot or a VCA and which kind of tasks it will perform.

– What channels will it be used on?

Once again, depending on whether you need an internal or external tools, the preferred communication channels will vary. Maybe your customers rather contact you via your website, or social media profiles, while internally you prefer software applications such as Teams or Slack.

– What kind of development are you looking for?

Do you look for a customized solution that responds to the specific needs of your customers/internal users? Or do you want a simpler, more generic tool that funnels the users towards a one-to-one interaction? In either case, you’ll need to describe an algorithm for the AI actions, develop an answers’ database and test the solution before launching it.


Nowadays, there are a bunch of different AI solutions on the market, that promise you enormous benefits if you develop your own DIY solution. However, beware of those smaller players that promise you’ll be able to develop your own Alexa as easily as you open up an email account. A DYI solution might be tempting, but it will end up being really expensive on the long run. Unless you are a very skilled coder, those DIY robots will be really hard to develop and, in the end, they won’t be able to meet customers’ expectations regarding the required level of assistance, damaging your brand reputation by offering a poor customer experience.



With SuperCards we are offering you an alternative to all those pains. SuperCards is not either a chatbot or a virtual assistant, it’s an integral solution that takes the best of both technologies and actually delivers real usable support for business, internally and externally.

Receive Our Posts In Your Email

Do you want to keep up to date with our blog without having to come and check what’s new? Let’s us help you!