Virtual Assistants

From New Media Business Blog

Jump to: navigation, search


Introduction: What are Virtual Assistants?

The term virtual assistant can cover a variety of different topics, and if you go to google and search for virtual assistant, you return 3.49 million different search results. [1] This is a large variety and can provide many different meanings. However, we will be focusing on virtual assistants that are not humans, aka machines. In the past, and still today, virtual assistants are available in the human form as outsourced employees that manage communications, emails, etc. If you continue to search in to virtual assistants, you will find more and more information regarding AI systems that help with a multitude of tasks.

History of Virtual Assistants

Figure 1 IBM engineer William C. Dersc demonstrating the Shoebox [1]

Speech Recognition Machine: IBM Shoebox

In 1961, IBM developed a device that was able to recognize and respond to 16 spoken words, being the first ever isolated digital speech recognition tool. IBM Shoebox was able to recognize the ten digits from “0” through “9” and command words such as “plus” “minus” and “total” for simple arithmetic problems. This experimental device was an early effort at voice recognition technology developed by William C. Dersc at IBM’s Advanced Systems Development Division Laboratory.[2]

Animated Office Assistant: Microsoft Clippy

Figure 2 Microsoft’s Clippy, an interactive animated character [2]

Developed by Microsoft in 1996, the office assistant was an intelligent user interface designed as an interactive animated character to help with Office content. Designed by Kevan J. Atteberry, Clippy drew criticism from users, employees and even Macintosh users. Clippy’s main purpose was to help users navigate using Microsoft applications like Word or Excel. Although its intent to help was good, Clippy was hated for the in-your-face annoying personality in combination with its terrible interface, Clippy was optimized for first time users displeasing any other user. [3]

Mobile (Smartphone) Voice Assistant

Today, almost all leading smartphones include a virtual voice assistant to help with day-to-day tasks. Apple, however, was the first company to introduce their voice assistant, Siri, on the iPhone 4S in October 4, 2011. Siri is a virtual assistant that is pre-programmed in to all iPhones and most Apple products including computers and tablets. This was seen as one of the first virtual assistants that was available widely to consumers. Originally released as an application on the iOS system available on the app store, it was purchased and integrated in to the iPhone itself. Its main focus was to aid in simple tasks such as sending a text message or making phone calls to people on the local device. Over time, it has developed to provide you recommendations to restaurants, search the internet and provide directions.

Smart Home Devices with Virtual Assistants

Amazon Alexa

Amazon has their own virtual assistant that is able to live wherever you want, as long as there is an internet connection. It resembles many Bluetooth speakers on the market, although it does have that functionality, it has many more capabilities. With the growth of the Internet of Things, if connected to other smart devices in your home such as the lightbulbs, you can ask Alexa to dim the lights. Alexa is always listening to the conversation in your home for the wake up word “Alexa”. Once it has heard the word, it listens to gather voice commands. Once you have asked Alexa a command, it renders the words to a voice recognition system that is stored in Amazon’s cloud. This service named Alexa Voice Service, will interpret what you have said using machine learning and sends back a matching answer. [4]

At UBC, computer science master’s students have been able to automate particular parts of software engineers’ jobs. This is using the same technology that allows us to ask Alexa simple requests such as the weather. It became useful to complete minimal and simple tasks. The software engineers, a group of 21 were asked to test the software and had many positive reviews. The only challenge was the fact that in order to utilize Alexa, it had to voice activated. This provided challenges in an office environment. [5]

Google Home

Google Assistant is a virtual assistant powered by artificial intelligence. The main goal said by Google’s CEO is to engage in “two-way conversation”. [6] The assistant has all the main capabilities that most other virtual assistant has, such as playing music and connecting to other smart devices in the home through IoT.

In May 2018, Google announced Duplex. Duplex is an extension of the assistant that allows it to speak as if it was using a human voice. The assistant has capabilities to autonomously complete tasks such a booking appointments, scheduling restaurant reservations. It uses more normal human speaking deficiencies such as “uhh” and “hmm” to make it sound like a human. [7]

The voice activated assistant begins with “okay Google” or “hey Google”. Over time, Google Assistant has been built to recognize different voices profiles to understand who is talking to it. It also understands connect to react in an informed and smart way.

Vehicle Virtual Assistants

Voice control first appeared in vehicles in the 1990’s, mainly used for elementary commands. Despite the large technological advances in virtual assistants that can be seen in smartphones and smart speakers today, the same cannot be said for voice assistants within vehicles. Improvements in speech recognition need to be made in order for a vehicle’s virtual assistant to be useful to the driver. This makes sense as a vehicle manufacturer’s priority when engineering a new vehicle are not usually on improving their voice assistants.

Figure 3 BMW Intelligent Personal Assistant [3]

However, Microsoft and BMW have partnered together to build a virtual assistant for the modern automobile. The automaker will begin optimizing their assistant to control the vehicle, whereas Microsoft’s aim will be to offer a higher level of productivity to its drivers. Temperature control, lighting, and voice commands will give BMW a competitive advantage over other automakers with this new technology. This new effort called the Open Manufacturing Platform is aimed at developing and encouraging collaborative IoT development, focusing on smart factory solutions and building standards to develop in areas like machine connectivity and on-premises systems integration. [8]

Virtual Office Assistants

As the rate of technology accelerates, the relationship between automation and unemployment rates are constantly discussed. Virtual Assistants (VA) are self-employed, providing professional administrative, technical and creative assistance to remote clients from a home office. These individual contractors are not responsible for employee-related taxes, insurance or benefits. Furthermore, clients avoid the logistical problem of providing office space, equipment or supplies, reducing costs and providing an opportunity to commit resources to other value adding activities. [9]

Some platforms where individuals can connect with virtual office assistants are the following:

  • Upwork [10] - virtual assistants that offer services in a range of different fields
  • Codeable [11] - platform that connects customers with skilled Wordpress freelancers
  • Guru [12] - a freelance marketplace for virtual assistants in various fields

Perceptions of Virtual Assistants

The Personal Digital Assistant has been adopted by many consumers rapidly, however virtual assistants has seen little real world usage. Research from Sandeep Chowdhury from Coventry University wrote about the avoidance attitudes towards virtual assistants, suggesting that interaction avoidance with a device is influenced by social norms and perceived ease-of-use and suggests that older adults likely avoid interacting with a device due to a lack of privacy. The study also found that people aged 38 and above had a significantly lower mean score on their perceived VA ease-of-use which signified attitudes determined by control difficulty between age groups. [13]

Figure 4 Technology Acceptance Model [4]

Figure 5 Technology Adoption Lifecycle (TAL) [5]

Adoption by Customer Segments

The Technology Adoption Life Cycle is a sociological graphical model that illustrates the adoption or acceptance of a new product or innovation according to five segments. Innovators are the first to adopt an innovation as they are willing to take risks and have the financial resources to help absorb the cost of adopting if it ends up failing. They are the youngest amongst the segments and are willing to take risks. In terms of virtual assistants, individuals in the early majority, late majority and laggard segments are not willing to interact with these devices due to lack of privacy. [14]

How They Work

Upon arrival, the words and tone of your request to your virtual assistant are analyzed by an algorithm, which are then matched with a command that the assistant thinks you asked. Essentially, it is saying “we are eighty-five percent sure you asked this question,” which is why you do not always get the answer you were looking for. If the algorithm is not certain enough, it may ask “did you mean this instead?” and give you its best guess as to what you wanted. Or it could say “I’m sorry I can’t do that yet.”

Natural language processing allows it to understand the command and connect to the appropriate services to provide an answer. This depends on whether or not this is a local question that the device can answer quickly or it needs to connect to the internet.

Depending on the type of request, local such as “Pause music” that is simple enough and can be done almost immediately. Depending on your internet connection, and if you it needs to connect to the network when you ask a question such as “What is the weather in LA” it might take a bit longer.

Alexa uses their own Amazon integrated cloud service to answer questions in to the device. [15] Alexa is also connected to your Amazon account so it stores and tracks data about your purchases and current habits.

Google Assistant is probably the most knowledgeable as it has the entire Google storage at its disposal to answer your questions.

Siri prides itself on the privacy of their service, as your Apple ID account is not associated with Siri but rather only local storage information such as contacts, Apple Music and your locations (for mapping purposes). Apple, for example, does not collect your photos on its servers. [16]

To summarize, AI is essentially a mix of machine learning which are sets of algorithms that are testing sets of data to find correlations and presenting that data depending on whether it is a local command or one that needs to connect to a network.

Types of Intelligence

Assisted Intelligence

AI automates basic repetitive and standardized tasks, working from clearly defined rules. Humans are still making the key decisions. Examples include automated assembly line robots as well as software-based agents that simulate the online activities of humans.

Augmented Intelligence

AI enhances the ability of humans to accomplish tasks, and humans and machines learn from each other. Examples include intelligent virtual assistants, some of generative design systems, and systems that can bring unusual or noteworthy events to human attention.

Autonomous Intelligence

AI takes over some decision-making, but only after a human fully trusts the machine or becomes a liability to the prompt accomplishment of a task. Self-driving vehicles, currently in development by over 30 companies, are just one example of autonomous intelligence.


According to the American Society of Mechanical Engineers, artificially intelligent systems will diminish the amount of time-intensive jobs that designers face such as fixing errors, determining optimal positions, finding content and communication changes. Machine learning will be able to take on those tasks from humans and complete them at a quicker rate. Moreover, AI will be to support in the creation of designs. Intelligent AI systems will be optimizing supply-chain processes and delivering them to manufacturing facilities. [17]


Figure 6 Asus Zenbo providing childminding services [6]


Zenbo is a virtual assistant that is active and mobile. Zenbo can be used as a front line guest services agent, to greet customers and take basic orders at places such as a hotel or fast food restaurants. Zenbo also functions as an early childhood educator, it works to interpret faces and actions of young children to teach them things such as ABC’s and numbers. Zenbo utilizes deep learning and AI to connect patterns regarding human emotion to deliver the most effective learning processes. [18]

Current State of Virtual Assistants

Virtual Assistants have grown significantly in popularity due to the increased accessibility and integrability of the technology. Promotions such as Spotify giving out free Google Home Minis to its premium users have broken down the barriers to purchase for virtual assistants. [19] Additionally, the growth in complementary and compatible smart devices has benefited Virtual Assistant sales.

North America is one of the largest markets for Virtual Assistants, with 42% of US adults reporting that they use their phone voice assistants regularly. [20] In Canada, 5.8 million Canadians, 15% of the Canadian population report using a smart speaker at least once a month. [21]

Figure 7 Smart Speaker Use Case Frequency January 2018 graph from Voicebot Smart Speaker Consumer Adoption Report January 2018 [7]

The picture above depicts common usages of Smart Speakers as of January 2018. The virtual assistants are most frequently used for basic tasks such as answering questions and playing music. [22] Making purchases and messaging people prove to be more complex tasks that users do not yet trust Virtual Assistants to perform correctly. However, the technology to perform these tasks accurately through the use of past purchase data and commonly purchased items is already being employed by Amazon’s Alexa and other smart speakers. [23] One of the biggest barriers to adoption for new technological advancements in voice assistants will be trust; the willingness to give up control over some of our day to day tasks for the sake of convenience.

Industry Leaders

North America

Despite the concern for privacy and trust in the technology, the market for virtual assistants has been rapidly growing. It is predicted that by 2025, 75% of US households will own a smart speaker. [24]

Figure 8 Microsoft Survey Results on Voice Assistant Usage [8]

In the North American market, Google and Apple currently lead, with Alexa following close by. Please note that the image does not depict actual usage; many of these virtual assistants come preinstalled on purchased devices. [25]


The Chinese virtual assistant market is experiencing tremendous growth. Many of the world’s leading technologies are being developed by what may be lesser-known brands outside of China, but are currently dominating their local market.

Globally, Baidu, Alibaba, and Xiaomi are experiencing rapid growth due to their control of the Chinese market. [26] While Amazon and Google are slowing down, China’s growth has just started.

The three companies fighting for the majority of the Chinese market are Baidu, Alibaba, and Xiaomi. They are currently at the forefront of virtual assistant technology, yet we may never see their products expand outside of China.


Baidu is one of the largest AI and technology companies in the world, boasting the second largest search engine in the world, among many other services and products. [27] They are leading AI research in China, resulting in voice assistants with high accuracy. The utilisation of their conversational AI system, natural language processing, and speech and image recognition rivals Google. [28]

Some of their innovative products are:

  • Baidu is taking a different approach with their AI technology. Their Duer OS intelligent chip turns anything compatible with it into a smart device. They sell this software to other developers and actively invite others to collaborate on their platform. Many companies such as HTC and Lenovo utilise Duer OS on their products. [29]
  • The Xiaodu TV Mate is a 3 in 1 smart speaker, 4K HD set-up box, and home theatre system. Some of the TVs more interesting features include their image recognition, which is able to recognize celebrities that appear on the screen. Additionally, their far-field voice recognition technology is able to activate from up to 5 metres away. [30]
  • The Xiaodu Smart Display 1S is a smart speaker with a screen. The Smart Display addresses the concerns of children using smart speakers by implementing a child-friendly mode. Kids mode on the speaker activates when it detects a child's voice; it then restricts control to protect from inappropriate content and responds in a child-friendly voice. [31]


Alibaba is a multinational conglomerate best known for its e-commerce platform, which dominates the Chinese e-commerce market. [32] Through their investments, they are one of the world’s largest internet and AI companies as well. [33] Their rapid growth and investment in new technology rivals Amazon.

Alibaba is focussing on technology based on where we spend the most time outside of work - in our cars, and in our bedrooms. Their basic speaker, the Alibaba Tmall Genie, is cheaper than Alexa, and almost half the price of the Google Home. [34]

Figure 9 Consumer using Alibaba’s smart mirror for makeup application [9]

Alibaba A.I. Labs recently released a smart mirror that has partnered with beauty companies to give beauty tips, branded content, and reminders. Additionally, the lighting can be changed to replicate natural sunlight or dimly lit environments to ensure makeup application is flawless. [35]

They have also developed an in-vehicle smart speaker, with the ability to connect to the car to provide directions and other assistance. Additionally, the speaker can attach to the cars dash camera to project directions and visuals onto the road itself by integrating augmented reality technologies. [36]


Xiaomi is about to overtake Apple as the third largest smartphone company, behind leaders Samsung and Huawei. [37] Xiaomi Corporation is a consumer electronics company who has been focussing heavily on developing the smart device ecosystem within the home. [38]

The Xiao AI Speaker is priced around $29USD and is meant to be at the center of all the internet-connected devices in the home. Through wi-fi, infrared, and Bluetooth, the speaker is able to connect to up to 5000 different devices. [39] Appliances such as rice cookers, fans, humidifiers and refrigerators will become very easily accessible.

The Xiao AI English Teaching Gadget will make learning English a lot more convenient. Priced at $72 USD, the gadget boasts five different dictionaries and teaches through listening, speaking and providing visuals. The device provides the correct scientific pronunciation of English words to 224 different countries. [40]

Societal Impact

Privacy and Ethics

With the growth of the Internet of Things and virtual assistants, cybersecurity and consumers’ privacy are more pertinent. In order for companies to improve their virtual assistants, specifically smart home devices, they need to understand how customers are using these devices and for what reason. According to their privacy policies and public statements, both Google and Amazon state that they do so by listening to and dictating their customer’s anonymized audio recordings from Google Home and Amazon Alexa devices, respectively. Employees of Google and Amazon listen to about 0.2% of all audio recordings. Because these recordings have been anonymized, employees are not about to access a particular user’s recordings, or the data collected by their smart device.

By listening to customers’ audio recordings, companies are able to enhance their customer’s experiences with their virtual assistants by:

  • Improving the device’s speech recognition to better understand different accents and languages
  • Enhancing ad personalization for each individual user

However, both companies’ privacy policies include many contradictions, making it unclear of what information is being collected and what these recordings are actually being used for. Despite the many statements that privacy is a top priority by North America’s industry leaders, these companies are creating controversial headlines regarding breaches and mismanagement in their customer’s data on a regular basis.

Information Collection

As more devices are created with the ability to connect to the internet, smart speakers and virtual assistants will become important fixtures in our connected household. Currently, smart speakers are able to collect data in our home through the usage of smart lights and smart locks. Google and Amazon both want to increase the amount of data they gather by requiring connected smart devices to send a continuous stream of data to the virtual assistant. As a result, voice commands will be processed faster, and everything will be easier to manage because it is up to date. However, both companies’ policies regarding privacy, and how they are using the data is not transparent. This may be the start of increased data collection in more private areas. Many companies have data regarding how we spend our day outside the house, but are limited in how we spend the time inside. Our household patterns, such as occupancy, when we get home, what channels we watch and in what room of our house could soon be used for marketing purposes. 25% of smart speakers are used to control something else, a number that is rising rapidly with the advancements in the Internet of Things technologies. Privacy and data collection concerns must be addressed as we give up more and more control. [41]

Figure 10 Survey on reasons why some consumers have not purchased a smart speaker [10]

Privacy versus Convenience

Despite the overwhelming privacy concerns that are associated with virtual assistants, the predicted growth of the virtual assistant industry begs the question of whether convenience is more important to people than their own privacy. Only 16% of respondents in a 2018 survey conducted by Voicebot.AI stated that the reason they did not have a smart speaker was a result from privacy concerns. [42]

The industry is realizing that consumers are not as concerned about their privacy as much as they claim to be. The success of the many promotions and deals for virtual assistants further reinforces this. In fact, the University of Washington conducted a report in 2017[43] where they found that people who did have concerns about their privacy and data being breached as a result of using virtual assistants did not actually do anything different in their everyday lives. Although consumers know about the potential security implications with owning a virtual assistant, the threats do not outweigh the convenience and want of having a device like Amazon Alexa or Google Home in their home. The University of Washington report stated the reasons why this is the case:

  • They have trust in companies to properly handle user
  • They do not feel like they are a “worthwhile” target
  • They have set up secure systems to minimize the chances of an attack

Gender Bias

When comparing leaders in the virtual assistant industry, there is one similarity among them all: their assistants are distinctly feminine by default. Apple’s Siri, Microsoft’s Cortana, and Amazon’s Alexa devices were designed to have a feminine voice as academic work shows that society prefers a female voice over a male one. [44] Although these companies do not explicitly state that their assistants are gendered, it makes it difficult to draw the line when consumers start to use the terms “she/he” rather than “it” when describing their voice assistants. This type of justification from companies contributes to the problem of gender bias. Calvin Lai, a Harvard University researcher states that “… the more that culture teaches people to equate women with assistants, the more real women will be seen as assistants – and penalized for not being assistant-like.” [45] Furthermore, the engineering teams developing these assistants are predominantly male and have programmed assistants such as Siri and Alexa to respond to verbal abuse with flirtation or deflection. In Quartz’ article, “Siri, Define Patriarchy” [46], they captured the different responses that voice assistants had when met with verbal sexual harassment.

Voice Assistant Responses to Verbal Sexual Harassment [11]

Gender Inequality in the Tech Industry

The International Telecommunication Union (ITU) estimates that only 6% of professional software developers are women. [47] However, this gender imbalance does not seem as wide as many people associate Alexa and Cortana, who are viewed female-gendered, as the face of Amazon and Cortana, respectively. With the growth of virtual assistants, if more and more companies are developing feminine assistants with engineering teams who are predominantly male, gender bias will become even more prevalent.

Future Outlook


The 2010 Gartner’s Hype Cycle for Emerging Technologies listed Virtual Assistants in the Trough of Disillusionment citing that at this stage “Interest wanes as experiments and implementations fail to deliver. Producers of the technology shake out or fail and investment only continues if the surviving providers improve their products to the satisfaction of early adopters.” [48] Apple’s Siri was launched in 2011, yielding lots of excitement from consumers about this innovative technology. We could see a natural progression of how VA’s will begin to shape our lives in the future. However, as we compare with the hype cycle in 2018, VA’s has taken a new position, in the Peak of Inflated Expectations. The prediction being in 2-3 years that VAs will be able to transition into mainstream adoption. Although we believed that Siri was an emerging technology that has many potential benefits, 10 years later we can see that nothing has really changed and we are still left in a phase where the true benefit of this technology is not actualized.

Figure 11 Gartner’s Emerging Technologies Hype Cycle, 2010 [12]
Figure 12 Gartner’s Emerging Technologies Hype Cycle, 2018 [13]

Figure 13 Scenario Planning: Virtual Assistants

Scenario Planning

Scenario planning is that part of strategic planning which relates to the tools and technologies for managing the uncertainties of the future. We can use this tool in order to predict what the future might turn out to look like at our current present day. We identify what trends we can see changing and the drivers causing change and list them according to uncertainty and importance. We rank the uncertainty based on how much uncertainty there is of the impact of that trend or driver and the importance of the impact of that trend or driver. We were able to identify 2 trends and drivers that were high in importance and uncertainty in order to create a 2x2 matrix. The New World is a hypothetical future outlook for our society where SMART technologies has been integrated with IoT devices with VAs having full capacity to operate these devices. Our world will become interconnected utilizing expansive 5G and possible 6G networks connecting IoT devices for seamless connectivity. VAs will be able to control devices and provide exceptional functionality for consumers. On the other end of the spectrum we have Stagnation, where VA technology has become obsolete and isolated and we are left with devices that cannot be controlled using VA technology. This is similar to the Age of Gadgets, where VAs have been integrated with IoT devices however these devices do not have cross functionality.


Nick Bucovaz Jennifer Dao Brian Kim Kyle Lee
Beedie School of Business
Simon Fraser University
Burnaby, BC, Canada
Beedie School of Business
Simon Fraser University
Burnaby, BC, Canada
Beedie School of Business
Simon Fraser University
Burnaby, BC, Canada
Beedie School of Business
Simon Fraser University
Burnaby, BC, Canada


  44. Stern, J. 2017. Alexa, Siri, Cortana: The problem with all- female digital assistants. Wall Street Journal, 21 February 2017
  45. Lai, C. and Mahzarin, B. 2018. The Psychology of Implicit Bias and the Prospect of Change. 31 January 2018. Cambridge, Mass., Harvard University.
Personal tools