Go to nearly any business website today, and you will likely run into a chatbot or intelligent virtual assistant (IVA). These chatbots have been built by several different software companies for businesses to support or sell you something. Perhaps the largest gathering of such bots today exists on the Facebook Messenger platform. Thanks to APIs provided by Facebook, which made creating and deploying agents simple, the number of live "messenger bots," as they are known, has eclipsed 100,000.
For businesses, the accuracy and efficiency of the IVA or chatbot directly impacts customer experience and dramatically reduces company support costs.
In one case study, Aetna saw a 29% reduction in live-chat volume within five months of deploying an IVA on their website. Domino's Pizza reported that product order time was reduced by 50% through their IVA. IVAs are so effective that analysts predict that human contact center agents will be replaced entirely by IVAs by 2026.
Already, the lines are blurring between virtual and human support agents. For example, Amtrak has reported customer confusion on whether or not its IVA for transportation services is a real person. This blurring of perception indicates that IVAs are presently capable of replacing humans in customer service for everyday tasks.
However, there remain many instances where users would prefer to have their needs handled by another human. These are cases when humans prefer a human touch, where sensitivity and compassion are required. For instance, when assisting with a deceased spouse's accounts or reporting improper employee conduct, human agents and their relational abilities are preferred.
The major misnomer is that in these special cases where users want a personal touch, human customer service agents often use canned responses or auto-populated templates to keep calls short. They simply click on a response or answer label, and the pre-defined response text is displayed to the user on the other end of the chat. In these instances, the end-user experience is no different than it would be in talking to an IVA or chatbot. However, unlike the IVA which has a lightning-fast response time thanks to natural language understanding and AI, the human agent still needs to read or listen to the customer's request and decide on which appropriate response to select.
Delivering high value with a productive, human touch
The ideal situation is to enable machines, like IVAs, and humans to work together to augment and maximize the effectiveness of both and deliver the best customer experience and outcome. In this situation, human customer service agents focus on the conversations that require specialized service, while machines scale to meet the large volume of redundant needs.
At Verint-Next IT, we have created a means to allow human agents and IVAs to seamlessly co-handle conversations with customers. This method allows IVAs to handle moderate or straightforward tasks, reserving humans for those interactions that require human compassion or special handling.
The traditional chatbot-customer support workflow starts with an IVA and, if errors occur or the need arises, the user is escalated to a human agent to complete. Or in many cases, the user skips the opportunity to try the IVA first. Our Collaborative Intelligence Model allows IVAs and human agents to work together on conversations, working together as a team to boost efficiency while at the same time delivering superior customer experience.
Here's how the best of both worlds collide
In the collaborative intelligence model, the IVA works alongside the customer service agent handling the initial greeting and user query. The IVA's language model contains intents that have been identified as those that must be handled by its human counterpart. For example, if the user asks to file a complaint about a recent experience with the company, the concept of "complaint" would flag the conversation to be continued by the human agent.
In the traditional workflow, the IVA responds to the user with something like "I'm sorry to hear that, please wait while I transfer you to a human representative." The user is then placed in a holding queue for the next available service agent. In the Collaborative Intelligence Model, the IVA passes control of the conversation over to its paired human agent. The human agent then responds directly to the user as if they were talking to a single agent. Once the user no longer requires human attention, the human agent returns control over to their IVA to continue the conversation seamlessly.
Other than a visual cue letting the user know when they are talking with the bot, and when they are talking with the human, the experience is seamless. And, unlike the traditional workflow, it's designed to more effectively deliver quick, responsive service; respecting the user's time and letting them get back to life.
Productivity is good for everyone
Having IVA's collaborate with human agents in the contact center brings about an entirely new wave of efficiency and higher productivity. Consumers are put on hold less frequently and for less time. Agents no longer have to scramble for several stressful minutes to find basic information, and managers are able to focus on creating more empathetic human connections with customers and workers alike.
At Verint-Next IT, we not only deliver voice and digital AI solutions for businesses to engage and support their customers, but some of our most powerful solutions have been supporting workforces and contact centers through real-time natural language interactions for over ten years. These AI-powered solutions shorten training times, improve response accuracy and compliance, accelerate time to insight, and decrease escalation calls to management. If you're interested in hearing more details on how AI-powered virtual assistants and human agents can work together in the contact center get in touch.