What to consider before adopting an intelligent virtual

We are excited to bring Transform 2022 back in-person July 19 and virtually July 20 – 28. Join AI and data leaders for insightful talks and exciting networking opportunities. Register today!


Contact centers have evolved to be dynamic communications hubs that have been put to the test the past two years. 

Companies have begun to invest in intelligent virtual assistants (IVAs) because they are effective in improving contact center productivity and the customer experience. However, to get the best return from these virtual assistants, you need to know your strategy. Without clear direction, you ultimately jeopardize customer experience. 

Here are questions to ask and challenges to consider before expanding your IVA strategy. Checking these boxes will help ensure the IVA meets your business needs and customer communications preferences.

Question: What level of complexity will the IVA support?

As I noted above, one of the first and most important questions you should ask is, “What is the general strategy for the IVA?” Is the IVA going to supplement your agents to allow for them to focus on more complex tasks? Or is the IVA going to focus on one or a few very specific use cases (e.g., password reset, bill payments or two-factor authentication)? 

When diving into your IVA strategy, it’s really about knowing the complexity you want the IVA to handle and how many of those inquiries you wish to block from being escalated to live agents. A clear strategy and knowing the complexities that could lie ahead are critical to successful integration. 

Challenge: Understanding the technology

Understanding the technology is central to designing IVAs that will support the required complexity. Knowing the differences between IVAs and other contact center solutions such as chatbots, voicebots and interactive voice response, known as IVR, will help you guarantee your IVA can effectively support specific use cases, regardless of complexity. Below are different contact center technologies and their key differences.

  • Chatbot: A chatbot is a program that can automatically communicate with a user without a human agent’s help. They have limited capabilities and typically interact via text. Chatbots are rule-based and task-specific, which allows them to pose questions based on predetermined options. They lack sophistication and  will not make any inferences from previous interactions with customers. Chatbots are best suited as a question and answer use cases. 
  • Voicebot: Voicebots and chatbots have similar functionality. The main difference from a chatbot and voicebot is the channel. Voicebots involve more complexity as they incorporate speech-to-text, which allows callers to speak to the bot. These solutions use IVR software.
  • IVR: Briefly mentioned above, IVR software is an automated phone system technology that interacts with callers and gathers information based on how the caller navigates a call menu. It does not use AI. Callers move through menu options through spoken responses or by pressing numbers on their phones. IVR software routes the caller to specific departments or specialists. Some may consider an IVR to be a simple voicebot.
  • IVA: An intelligent virtual assistant is the most sophisticated of the options and you can use it across various channels. IVAs process natural language requests using natural language understanding or natural language processing and understand situational context, allowing them to handle a more complex range of questions and interactions. These tools closely resemble human speech and can understand queries with spelling and grammatical errors, slang or another potentially confusing language, much like a human agent.

You’re better equipped to advance existing contact center communications strategies when you understand IVAs, the full volume of capabilities they offer and how they differ from other AI-enabled solutions. 

Question: What persona should the intelligent virtual assistant represent?

For an IVA to be effective, you must understand the persona you want the virtual assistant to represent. This persona will inform how you design your virtual assistant to act based on your company’s brand. To know the persona, you need to know how your customers engage with the contact center and the complexity of the skills that the assistants — live and virtual — need to be able to manage. 

Based on these defining characteristics, you can set business rules for the IVA. These rules then create the standard for how to design the IVA. Key questions to answer to uncover persona include:

  • Should the voice be female? Male?
  • Should it have an accent?
  • How many languages should it be able to speak?
  • Will it need to be familiar with jargon from a particular industry?
  • Should it have a casual tone and follow a more informal language model? Or should it be formal and professional?
  • How will customers speak to the IVA?

Answering these questions will guide you in designing an effective IVA that you can scale for your brand.

Challenge: Lack of collaboration between IT and CX teams

IT teams often work closely with a communications provider to design and implement the IVA. Though they support this process, IT teams typically don’t engage with the customers and might not have a clear picture of their engagement preferences. You can overcome this challenge by increasing collaboration between IT and customer experience (CX) teams.

For example, CX team members can provide insight into the company rules for customer support and how the business manages interaction paths and escalation levels. In banking, this might include the ability for a caller to create a payment plan with an IVA over the phone; however, if the IVA hears a specific balance figure or concern through a particular phrase, it knows to connect the caller to a human agent. If the IVA doesn’t have this level of business logic, the company can jeopardize the customer experience.

CX team members are also knowledgeable about how to create personas for customers and how to understand their engagement preferences. They’re also aware of standard industry terms that customers might use when interacting with an IVA that the IT team might not consider. Once IT teams know these terms, they can then create training models for the IVA that include the common terms and phrases.

What the future holds for intelligent virtual assistants

One current limitation of IVAs is that they sometimes lack visual engagement. It will be interesting to see IVAs evolve to video channels in the coming years. With video, customer support teams, through the use of IVAs, would use biometrics to understand people’s body language and experience, make inferences about their experience and sentiment and automate video support experiences or escalate to an agent. 

For example, in healthcare settings, if someone with a severe illness called their doctor’s office and communicated via IVA-enabled video, the IVA could visually pick up on common symptoms the patient demonstrates. This might include lack of focus, inability to maintain eye contact, drowsiness, etc. The IVA can then note these visible symptoms in the patient’s chart to inform the team of nurses and doctors. The potential of this technology is exciting.

Answering essential questions and addressing challenges related to using IVAs early in the investment process will help you optimize your strategies to leverage automated and intelligent solutions that improve customer experiences. As you deepen your IVA strategies, you’ll better understand the technology’s potential, improve customer experiences and see positive impacts on your operations.

Tim Wurth is director of product management at Intrado.

DataDecisionMakers

Welcome to the VentureBeat community!

DataDecisionMakers is where experts, including the technical people doing data work, can share data-related insights and innovation.

If you want to read about cutting-edge ideas and up-to-date information, best practices, and the future of data and data tech, join us at DataDecisionMakers.

You might even consider contributing an article of your own!

Read More From DataDecisionMakers