Intelligent virtual assistants (IVAs) are systems that use artificial intelligence (AI), machine learning, natural language processing (NLP) and natural language understanding (NLU) to automate the handling of customers’ and employees’ requests and tasks that can be addressed and resolved without requiring the cognitive capabilities of human beings. For customers and, increasingly, employees, IVAs can automate routine tasks and can process and personalise the handling of interactions in most self-service channels.
After more than 30 years of offering interactive voice response systems (IVRs) that required customers to think like a computer and work their way through nested choices, there is finally an option that addresses the need for improved voice self-service. (IVAs have also begun to be used to improve employee productivity and to enhance their satisfaction by making their jobs more interesting.
DMG Consulting research has confirmed that consumers in all generations, in many countries around the world prefer to use self-service solutions, as long as they work efficiently. Many of today’s IVR solutions are excellent, and traditional solutions perform well. The challenge is that these IVRs only work in one way, and what works for one person will not necessarily satisfy another. As a result, even some of the best-designed IVRs are not automating as many calls as they would if their self-service capabilities systematically adapted to the needs of each caller.
Justifying an Investment in an IVA
If you have not updated your IVR solution in the past 3-5 years and/or you’re not using NLP, it’s likely that you’ll be able to cost-justify an IVA investment with a 2- to 4-year return on investment (ROI) based on the scale and complexity of the initiative. (Your payback period will depend on the volume of activity and the channels to which IVAs are being applied.) The challenge is to present the financial proposition for the acquisition in a manner aligned with the investment policies of your organisation. For this reason, DMG recommends that organisations use a phased approach to transition to an IVA. This way you can show positive results before you have committed too many resources.
Implementing an IVA
To speed up approval for this essential investment, identify 1-3 activities that are fully or partially handled by agents, are high-volume, and usually do not require human intervention. An example could be handling of a consumer’s insurance benefits or claim processing status. In both situations, the caller wants only certain information and does not want to listen to many choices or go through many steps just to get the answer they need. (This is often the reason why customers request a live agent.) In the world of IVA, people who call, SMS, or send an email will be able to ask a specific question and rapidly get the answer they need. This will dramatically increase the use of self-service solutions, become the primary mechanism for helping customers, as well as give companies new ways to improve productivity.
The current generation of IVAs, utilising AI, machine learning, NLU, NLP as well as other types of technology, are ready for prime time. The vendors are responsible for the majority of today’s implementations and are developing best practices to improve the success and speed at which an IVA can be up and running. Vendors are helping users learn how to utilise these innovative solutions to improve and standardise their self-service capabilities.