Virtual Assistants
Enriching customer, employee, and agent experiences through a unique combination of conversational and digital UIs.
Defining a Virtual Assistant
A virtual assistant (VA) is an advanced computer program that can understand, process, learn from, and respond to voice or text inputs in natural ways. It can enable conversational interfaces with or without buttons or menus, and can understand and respond to complex user utterances. It typically combines advanced natural language processing (NLP), natural language understanding (NLU), and dialog management techniques to achieve a high degree of automation without sacrificing accuracy.
Key Traits of a Virtual Assistant
Simply put, software is complex, devices are small, and UIs often unintuitive. People don’t always know how to use or find specific features. Virtual assistants replace complicated UIs with natural language, so there is no learning curve and wasted time. And they can interact with almost any software that provides an API, so the scope of use cases is endless.
Virtual Assistants for Our Daily Life
When it comes to what virtual assistants can do, most think of simple use cases, like ordering pizza or telling you the weather. But in reality, the possibilities are endless. In fact, virtual assistants can handle most tasks, workflows, and service interactions with ease using chat and voice.
Enhance Customer Experience
A global retail bank was overwhelmed by requests related to account balances, card rewards, risk management, and more. In response, the bank deployed virtual assistants to over seven million cardholders across six business lines and multiple channels, greatly reducing call volume and enhancing customer experience.
Improve Customer Satisfaction
A leading manufacturing company was inundated with installation-related requests, leading to long call wait times. To reduce wait times and improve customer satisfaction, the company deployed a multi-channel virtual assistant to guide users through installation, thus slashing related call times from 15 minutes to 80 seconds.
Potential Use Cases
Banking
Insurance
Healthcare
Retail
Reduce Response Times
A multinational company was overwhelmed by password reset, account unlock, and equipment return requests. In response, the company built a virtual assistant that reduced task completion times from 24 hours to 40 seconds across 26 systems, while also reducing IT support load by 90%.
Improve Productivity
A global CPG company was struggling to process hundreds of transport management emails per day. Using virtual assistants, the company was able to automatically scrutinize, sort, and divert emails to the appropriate functional group, reducing $200K in annual costs and improving productivity 15%.
Potential Use Cases
ITSM
Sales
Accounting
HR
Increase Call Containment
A global telecom company wanted to create an engaging and effective IVR experience. With voice-based virtual assistants, the company automated 34% of existing intents, significantly improving call containment rates and reducing service costs.
Improve Agent Responsiveness
A multinational company’s agents were struggling to provide callers with answers in real-time. In response, the company built and launched a virtual assistant to help agents retrieve customer data from multiple backend systems fast – reducing retrieval time from 10 minutes to 2 seconds and average handling time by 240x.
Potential Use Cases
Call Automation
Agent Assistants
Agent Transfer
- Customers
Enhance Customer Experience
A global retail bank was overwhelmed by requests related to account balances, card rewards, risk management, and more. In response, the bank deployed virtual assistants to over seven million cardholders across six business lines and multiple channels, greatly reducing call volume and enhancing customer experience.
Improve Customer Satisfaction
A leading manufacturing company was inundated with installation-related requests, leading to long call wait times. To reduce wait times and improve customer satisfaction, the company deployed a multi-channel virtual assistant to guide users through installation, thus slashing related call times from 15 minutes to 80 seconds.
Potential Use Cases
Banking
Insurance
Healthcare
Retail
- Employees
Reduce Response Times
A multinational company was overwhelmed by password reset, account unlock, and equipment return requests. In response, the company built a virtual assistant that reduced task completion times from 24 hours to 40 seconds across 26 systems, while also reducing IT support load by 90%.
Improve Productivity
A global CPG company was struggling to process hundreds of transport management emails per day. Using virtual assistants, the company was able to automatically scrutinize, sort, and divert emails to the appropriate functional group, reducing $200K in annual costs and improving productivity 15%.
Potential Use Cases
ITSM
Sales
Accounting
HR
- Agents
Increase Call Containment
A global telecom company wanted to create an engaging and effective IVR experience. With voice-based virtual assistants, the company automated 34% of existing intents, significantly improving call containment rates and reducing service costs.
Improve Agent Responsiveness
A multinational company’s agents were struggling to provide callers with answers in real-time. In response, the company built and launched a virtual assistant to help agents retrieve customer data from multiple backend systems fast – reducing retrieval time from 10 minutes to 2 seconds and average handling time by 240x.
Potential Use Cases
Call Automation
Agent Assistants
Agent Transfer
Unraveling the Functionalities of Virtual Assistants
Virtual assistants use a combination of NLP, machine learning (ML), speech recognition, and other technologies to interpret a user’s words, match it to what they want, and generate an appropriate response. Similar to apps, virtual assistants include an application layer, a database, APIs, and a conversational user interface (CUI), and are available through platforms like Facebook Messenger, Slack, and others.
Conversational Interfaces
Until now, computers relied on highly-detailed, binary-coded instruction. Conversational user interfaces, which are made possible through NLP, enable users to “talk” to computers like they talk to friends, eliminating the need for computer-specific languages, icons, and syntax-specific commands. There are three types of CUIs:
- Voice assistants with whom you talk
- Virtual assistants with whom you text
- Hybrid experiences that meld conversation with digital elements
Most notably, a single virtual assistant build can include all three types of CUIs, with the type used dictated by the channel that the interaction takes place in.
Digital Interfaces
Digital interfaces traditionally consist of an amalgamation of virtual, auditory, and functional components. With virtual assistants, digital elements usually entail widgets. These components proactively display information and trigger functions, and forms, which provide input fields required for capturing details. By combining CUIs and digital elements, virtual assistants can provide engaging experiences, shorten response times, and reduce steps needed to complete processes.
While conversational interfaces aren’t new, technology has recently advanced enough to make everyday use practical. And, thanks to recent developments, it is becoming easier for computers to understand natural language – putting the onus on the software, and not the user, to figure out what to do and how to get it done.
Digital Interfaces
Digital interfaces traditionally consist of an amalgamation of virtual, auditory, and functional components. With virtual assistants, digital elements usually entail widgets. These components proactively display information and trigger functions, and forms, which provide input fields required for capturing details. By combining CUIs and digital elements, virtual assistants can provide engaging experiences, shorten response times, and reduce steps needed to complete processes.
While conversational interfaces aren’t new, technology has recently advanced enough to make everyday use practical. And, thanks to recent developments, it is becoming easier for computers to understand natural language – putting the onus on the software, and not the user, to figure out what to do and how to get it done.
Natural Language Processing
Natural language processing enables virtual assistants to understand language as people do. It can be broken down into natural language understanding (NLU), which accounts for complex facets like synonyms, slang, and mispronunciations, and natural language generation (NLG), which creates words and phrases in response to user input. With these, virtual assistants can identify intents and entities within user utterances, look for statistically significant patterns, and respond accordingly.
Channels
Channels are the medium through which users interact with virtual assistants. Omnichannel refers to the ability to deploy virtual assistants across multiple channels from a single build while providing a consistent and cohesive experience and taking advantage of the unique features of each channel. For example, some channels support images and videos, while others have gamified elements – all of which can be offered within a dialog. It also refers to the ability for users to start a conversation in one channel and complete it in another – all without losing history or context.
Channels
Channels are the medium through which users interact with virtual assistants. Omnichannel refers to the ability to deploy virtual assistants across multiple channels from a single build while providing a consistent and cohesive experience and taking advantage of the unique features of each channel. For example, some channels support images and videos, while others have gamified elements – all of which can be offered within a dialog. It also refers to the ability for users to start a conversation in one channel and complete it in another – all without losing history or context.
Task and Dialog Execution
While channels are the places where users and virtual assistants interact, tasks and dialogs are the methods through which things get done. Virtual assistants can collect, modify, and post information to systems of record, answer questions, retrieve reports and information from backend systems, send notifications and alerts, and more. Further unlike traditional apps, responses are not limited to raw inputs and outputs but can include or be influenced by personal data and past usage history, among other factors.
Training
For virtual assistants to get things done, they must be trained to recognize the combination of words typically indicating an intent. While this training can be done manually, the advantage of machine learning and other forms of training is that they allow computers to learn from experience without having to explicitly programmed by a human – significantly reducing the time needed to produce effective language models.
Training
For virtual assistants to get things done, they must be trained to recognize the combination of words typically indicating an intent. While this training can be done manually, the advantage of machine learning and other forms of training is that they allow computers to learn from experience without having to explicitly programmed by a human – significantly reducing the time needed to produce effective language models.
Conversational Intelligence
Human language is anything but simple. While training can go a long way, complex rules and nuance can quickly and fundamentally change the meaning and importance of a user utterance. To handle this challenge, virtual assistants are equipped with a number of capabilities allowing them to process complex utterances.
The Human Element
Today’s virtual assistants are smarter, more responsive, and more useful than the chatbots of old. Nevertheless, there will always be cases where a human is needed. Thus, as required, conversations can be transferred to live chat and voice agents, along with a full history with relevant context, user sentiment, request priority, user preferences, and more.
The reverse is also true – sometimes agents and reps could use a little backup. Virtual assistants can work in the background – in real-time – to provide support in the form of intent discovery, knowledge injection, script production, and more, allowing them to focus on the customer.
The Human Element
Today’s virtual assistants are smarter, more responsive, and more useful than the chatbots of old. Nevertheless, there will always be cases where a human is needed. Thus, as required, conversations can be transferred to live chat and voice agents, along with a full history with relevant context, user sentiment, request priority, user preferences, and more.
The reverse is also true – sometimes agents and reps could use a little backup. Virtual assistants can work in the background – in real-time – to provide support in the form of intent discovery, knowledge injection, script production, and more, allowing them to focus on the customer.
Putting It All Together
The technology that forms the bedrock of virtual assistants has evolved greatly and is finally sophisticated enough to tackle complex enterprise use cases. Thanks to virtual assistant platforms, building, deploying, and managing novel customer, employee, and agent experiences is easier than ever, requiring little to no coding or expensive software.
Business Benefits of Virtual Assistant
What Does This Mean for My Company?
Demand is growing for virtual assistants in the consumer, social, and enterprise world, and they are becoming more commonplace. Customers and employees increasingly expect them, and your competitors are probably considering them. Failure to innovate and keep up with changing expectations can put your company at risk. Don’t get left behind.
How Do I Get Started?
Let us show you. Either contact us to get your questions answered, or learn more about the virtual assistant journey.
