A version of this article first appeared as a guest post on G2's Learn Hub.
Intelligent virtual assistants and virtual agents for improved CX
In today’s digital-first world, where customer expectations continue to rise, more and more companies are turning to automation to help scale their teams with technology. Advances in artificial intelligence (AI) capabilities have made it easy and cost-effective for organizations of all sizes to adopt AI applications. One such application that’s gained enormous traction recently is the Intelligent Virtual Assistant.
Intelligent Virtual Assistants (IVAs), sometimes also known as virtual agents, are a type of specialized AI that is designed to be able to communicate in a human-like manner. They can understand the way people speak and can respond to all sorts of inquiries in a way that mimics actual human conversations. IVAs are often built for a specific purpose, such as providing customer service or qualifying leads, and they allow companies to scale almost instantly, while ensuring a great customer experience.
But what exactly are IVAs? How do they work? Will they really help your business to operate more efficiently? And is there any risk of them taking jobs away from humans? Read on for an in-depth look at everything you need to know about IVAs.
What is an intelligent virtual assistant (IVA)?
Intelligent virtual assistants are AI-powered software that can engage with a business’s customers in a conversational manner. Essentially, IVAs are ultra-sophisticated versions of chatbots that understand and can respond to natural language without being explicitly programmed to do so.
Think of them as a digital assistant that emulates human interaction to perform certain tasks, especially simple, repetitive tasks.
IVAs are used extensively in customer support automation, for example, to automate email, chat or social media conversations with customers. An intelligent virtual assistant can act as the first line of defence for your organization’s customer service team, then escalate complex cases to your human agents when necessary. By automating customer interactions, IVAs enable 24/7, omnichannel support for the customer while eliminating repetitive and time-consuming processes for your company.
While these bots replace the workload of many agents, they’ll never truly take the place of human support. The most efficient and effective teams operate with a combination of human and artificial intelligence to provide the best customer service. The IVA acts as a virtual agent that works with your live agents to field customer inquiries. The virtual agent can deal with any routine, high volume inquiries that would otherwise bog down live customer service agents, and the agents can use their human touch and customer service expertise to solve more difficult queries.
Not only does this reduce costs (since you’re paying for one virtual agent rather than hiring and training multiple human agents), IVAs have actually been proven to improve job satisfaction among live agents. Their workload is lightened by the IVA, and they feel empowered to work on complex and rewarding cases. In some cases, there is even the opportunity for them to become specialists in Conversational AI and bot building and training. This both reduces churn among employees and raises the profiles of customer service teams within companies.
Overall, intelligent virtual assistants and virtual agents can dramatically change how your business engages with its customers. IVAs allow businesses to streamline processes, quickly scale their teams, and positively impact both customer and agent satisfaction.
Intelligent virtual assistants vs. chatbots: What’s the difference?
IVAs and chatbots are often used synonymously, but IVAs are actually far more sophisticated. While both imitate human dialogue, chatbots are limited by a given script, whereas IVAs can engage in context-related conversations.
There are 3 main ways that IVAs and chatbots differ: how they speak to customers, how they provide answers, and how they learn.
How IVAs vs. chatbots speak to customers
Chatbots are programmed using rule-based algorithms. This means they are able to recognize certain keywords or a set of predefined questions and provide customers with answers to those specific questions.
IVAs, however, use neural networks and machine-learning-based algorithms to recognize what the customer is asking and respond accordingly. Their natural language processing (NLP) capabilities are more sophisticated in both input and output. That is to say, when a customer asks a question, IVAs can recognize synonyms, misspelled words, plurals, and informal language, and when the IVA responds, it’s able to give precise answers, even if it has not been explicitly programmed to respond to the exact inquiry at hand.
How IVAs vs. chatbots provide answers
Most chatbots are basic question and answer machines. Essentially, the customer must write the exact keyword required to receive the appropriate answer. They use scripted responses or direct customers to FAQ pages. If they are unable to provide an immediate answer, there is usually a way to escalate the inquiry to a live agent.
IVAs use conversational AI to communicate in a human-like manner. A combination of natural language processing (NLP), natural language understanding (NLU), machine learning, and AI help them to understand a customer’s intent and respond in a way that feels natural. IVAs are often integrated into an organization’s backend systems so they are able to use existing data or prior interactions to provide context for both the customers and the employees. This also allows IVAs to provide totally personalized responses. If and when necessary, IVAs can loop in agents to take on difficult to answer questions or queries.
How IVAs vs. chatbots learn
Chatbots are rule-based systems, so the only way for them to “learn” is by having a human create more rules. They do not retain information or learn how to better answer questions based on context.
IVAs that use deep learning models are able to learn from examples. This means that the more data you provide for your AI model, the better it gets at recognizing what the customer wants and responding to it accurately. The most advanced IVAs are also able to be trained. Human agents can delve into conversations that the virtual assistant did not understand correctly and train it to recognize the intent the next time it encounters a similar conversation. This allows the IVA to become particularly efficient and handle more inquiries as time goes on.
Learn how Ultimate's new Training Center makes your AI more accurate using historical support data and NLP.
How do intelligent virtual assistants work?
IVAs work using a combination of different AI technologies including deep learning, natural language processing (NLP), and natural language understanding (NLU).
Deep learning is a type of machine learning that allows the IVA to get better with every interaction and retain contextual information. NLP enables the IVA to understand the text including grammar, structure, and keywords. This differs from NLU, which allows the IVA to infer the intent behind the language used to derive meaning and context. Altogether, these technologies combine to power an intelligent virtual assistant that can predict a customer’s intent and provide an appropriate answer the same way a live customer service agent could.
But even with the most sophisticated AI technology, IVAs cannot predict what information users are looking for, so it’s important to ensure the data powering your IVA is the right content. Some IVAs use industry standard sets of data, but the best ones use your own company’s historical data.
Basically, every single customer support conversation that has occurred over the years is stored in your CRM platform. By uploading this data to your intelligent virtual assistant, the AI can cluster the most common topics together and thereby identify frequent customer “intents”. Once you’ve deciphered what your customers are most frequently looking for, you can create conversation flows to automate those types of inquiries.
When this process is based on your company’s historical data rather than pre-built data sets, you can be sure you’re addressing exactly what your customers need and customize your IVA’s dialogues to your organization’s brand personality and processes.
Another thing the more advanced IVAs are capable of is learning and improving. A “set it and forget it” model is often easier to set up, but will result in poorer performance in the long run. When you can train your IVA to recognize new intents, it means it will continue to become more and more efficient.
Where are intelligent virtual assistants used?
Intelligent virtual assistants are used across many industries, including ecommerce, fintech, travel, healthcare, and more. They can be especially useful for customer service teams in businesses that deal with a high volume of repetitive inquiries.
Using an IVA can also benefit companies that are particularly susceptible to rapid or unpredictable increases in support volume. For example, an ecommerce company that experiences huge increases in support tickets around Black Friday can make use of an IVA to field routine inquiries, freeing up the human agents to work on what really matters.
Rapidly scaling businesses can also make use of IVAs if it becomes too time consuming to hire human resources (or if hiring human agents cannot provide strong enough ROI).
In general, the types of companies that use IVAs tend to be B2C businesses that have existing contact centers that they can integrate into the IVA platform they choose.
No matter the industry, some reasons for implementing an IVA might be:
- To scale the company’s support team capacity
- To better serve their customer base with immediate support available 24/7
- To increase customer satisfaction score (CSAT)
- To drive more revenue
- To reduce costs related to employee churn
- To empower their customer support teams to work on challenging escalations instead of mundane tasks
The ultimate goal is always the same: to provide excellent support that keeps their customers satisfied while automating processes that help keep employees happy and save the business money.
Read more on the benefits of automation
5 benefits of intelligent virtual assistants
Introducing intelligent virtual agents allows businesses to dramatically improve their customer service processes. Here are some of the ways IVAs benefit both the customer and the company.
Deliver exceptional customer experiences
Today’s customers expect and demand efficient and personalized service. In a recent survey of top CMOs, 71% of respondents agreed that buyers expect an experience that is personalized to their needs and preferences, and 74% agree that buyers expect an immediate response to their questions.
By implementing an IVA that uses backend integrations, not only can you deliver quick answers, you can ensure those answers are completely personalized to your customers’ data and preferences.
IVAs allow customer support teams to save a significant amount of time and effort addressing routine inquiries. By using an IVA as a first line of defence, customer requests can be deflected from phone, email, and live chat channels, support tickets can be tagged automatically which reduces backlog, and agent time can be used on cases that require human intervention. This decreases wait times for everyone as well as saving costs on hiring and training live agents.
Improve your agents’ KPIs
Customer Satisfaction Score (CSAT), Average Handling Time (AHT), First Response Time (FRT), Net Promoter Score (NPS). What do all these acronyms have in common? They’re frequently used to evaluate customer service teams’ performance.
Companies that implement IVAs often see an improvement in all of the above metrics. Intelligent virtual assistants are able to either resolve inquiries or escalate the problem to an agent with the appropriate tags and context for a person to be able to address the issue faster, resulting in lower AHT. Bots can greet customers instantly, resulting in extremely fast FRT. And since IVAs increase efficiency overall, CSAT and NPS tend to increase, especially as time goes on and the IVA becomes better and better.
Reduce agent frustration & churn
Many agents in traditional contact centers spend a lot of their time handling mundane requests that follow similar processes, such as order tracking, returns, cancellations, etc. This can be frustrating for people who would like to handle more complex problems and perpetuates the age-old stereotype of call centers and customer service being a temporary job. By letting an IVA handle these routine tasks, like checking dates, names, and order numbers, your live agents can focus on harder to solve issues. In turn, this results in customer service becoming a more rewarding and sought-after career field, reducing attrition rates among customer support teams.
Provide global, 24/7 customer support
One huge benefit of IVAs is that they can operate constantly. For companies that operate across time zones or around the world, this is an invaluable aspect of virtual assistants. Instead of hiring people around the world, the IVA can provide round-the-clock support and escalation can be handled during regular business hours, drastically reducing backlog and agent workload, while providing customers with some level of service at all times. What’s more, some IVAs even have the capability to deliver multilingual support. This enables businesses to operate in multiple markets without having to hire as many bilingual or multilingual agents as they would without the IVA’s support.
Ready to start automating? Check out the 10 checkboxes to look for when evaluating the right IVA for your business.
Examples of intelligent virtual assistants
IVAs have been used to make an impact across various industries. Here are a few of our customer stories where Ultimate's intelligent virtual agent enabled better customer service:
How Finnair’s IVA helped them navigate the Covid-19 pandemic
In March of 2020, Finnair’s query volumes surged. Faced with an increase of 600 to 900%, Finnair needed to respond immediately to an influx of customers reaching out about cancellations and refunds. Thankfully, they had already implemented Ultimate's IVA for chat automation. By tweaking the welcome message to determine if the query was related to the current health situation and the ever-changing travel restrictions, the IVA could direct customers accordingly. They also implemented a new support path exclusively for refunds and cancellations in both Finnish and English. This resulted in an automation rate of 50% in March of 2020.
How Pockit's virtual agent reduced its email backlog by 95%
UK fintech company Pockit headed into 2021 with a massive email backlog and response times spanning several weeks. By adding an IVA to their support team, they were able to automate over 12,000 emails a month. Overall, Ultimate’s automated solutions produced 3.5 times more output than a human agent while implementation costs remained below one monthly salary.
How Telia saves 2,835 hours of agent effort a month
Telia, a Swedish-Finnish multinational telecommunications corporation, was seeing chat growing quickly — with up to 75,000 customer service chats occurring per month in Finland alone. After launching their intelligent virtual agent, Telia was able to automate 30% of all chats, saving 2,835 hours of agent effort a month (over 34,000 hours a year) during peak season thanks to automation.
A final word on intelligent virtual assistants
As more and more business is being done online, and teams are looking to scale quickly, automated customer support is quickly becoming the go-to solution. In fact, in a recent Deloitte Digital survey, 79% of contact center leaders said they plan to invest in greater AI capabilities in the next two years.
Industry leaders in travel, e-commerce, fintech and beyond have already greatly benefited from IVAs and will continue to do so. As automation becomes the norm, customers will become more accustomed to it, and companies that don’t adopt AI will be abandoned for more innovative companies that provide better customer experiences.