There is a lot of excitement on what generative AI can do for chat but it is not yet the channel of preference. When people want service, they still pick up the phone. Interactive Voice Response (IVR) technology remains the most important connection between companies and customers in many sectors, accounting for twice as many interactions as calls with live agents and five times more than text-based chat—and demand for these services is still growing. The global market for IVR systems is expected to reach $9.2 billion by 2030, up from an estimated $4.9 billion in 2022.1
Recognizing that IVR is here to stay, some leading companies are now giving automated voice channels the attention they deserve by applying smart-interaction design principles, advanced analytics, and deep-learning technologies to create systems that meet the needs of 21st-century customers.
The results can be spectacular. We have seen next-generation IVR systems deliver a fivefold improvement in customer satisfaction scores, accelerate issue resolution, and reduce the number of live-agent calls by more than 10 percent.
Yet too many companies still treat IVR as an afterthought, relying on outdated technologies and customer journeys that have changed little since the 1990s. Nowadays, customers want their issues resolved as quickly as possible with very little effort. Old-style IVRs require a lot of effort and don’t always provide a solution. The result: irritated customers, poor performance, and higher costs. Across industries, seven out of ten companies report that their containment rate, a measure of the percentage of calls that remain within the IVR system, is 30 percent or less.
This time it’s personal
Most companies employ non-personalized IVR systems, where all customers are treated the same, regardless of the information available about them. The key difference between high-performing IVR systems and their predecessors that remain in use is their customer-focused design. Rather than attempting to build a one-size-fits-all, general-purpose service channel, leading companies are tailoring their systems to meet the needs of individual users. They do that by taking advantage of these three key enablers:
- Enhanced IVR design. Applying human-interaction design principles to IVR systems helps companies make systems that are easier and less frustrating to use. This can involve changing call flow based on information provided by the caller or changing the way questions are worded for different contexts. Small details matter, too, such as the use of voice synthesis systems that offer an empathetic tone.
- Predictive engines. With machine learning technologies, companies can use data from customers’ on-call and historical actions to predict their likely needs and tailor the IVR journey accordingly. If a customer always calls the bank at the end of the month to confirm that a paycheck has cleared, for example, the system might offer a shortcut to do so on the first IVR menu.
- Next-generation technologies. The most advanced organizations are now building conversational AI and natural-language-understanding (NLU) technologies into their IVR systems. That’s paving the way for humanlike conversations with an automated agent that understands the customer’s intent and can offer a speedy and friendly resolution.
How much will it cost?
Crucially, smarter IVR doesn’t always need a major investment in new core technology. Most companies can make their existing systems work better with the addition of a few carefully chosen components. That might include a robust repository to collect and store data on interactions across different channels and an analytics platform to extract insights on customer behavior. Best-in-class organizations tend to follow a sequential or “wave” approach to the improvement of their IVR systems:
- Wave 1 uses advanced analytics to determine the “break points” in the current call flow, where customers give up or seek help from an agent. This gives companies a list of potential adjustments to the system and allows them to implement quick design changes to tackle some of the top sources of user frustration.
- Wave 2 redesigns customer journeys for the priority call types identified in Wave 1. This wave allows companies to apply user-centric design principles from a clean sheet. It’s also an opportunity to integrate available new technologies, such as call flows that adapt based on context data from customer activity outside the IVR.
- Wave 3 applies machine learning and AI technologies to add predictive and conversational capabilities to the IVR system.
Companies that want to move fast can conduct some of these activities in parallel. If the organization already knows which AI and machine learning (ML) platforms it is going to use in Wave 3, for example, it can start their implementation earlier, allowing Wave 2 activities to benefit from more advanced technologies.
What companies should not do, however, is skip Wave 1. It’s easy to get excited about new technologies without really understanding what makes the current IVR system frustrating for users, but that can embed bad design decisions into the new high-tech system.
The best companies take the opposite approach. They regularly and continually repeat their analysis of customer behavior, tracking improvements and updating their list of priority initiatives. Run that way, an IVR system can never get stuck in the past. It is always evolving and progressing in line with advances in technology and changes in user needs. In future posts on this topic, we’ll explain just what it takes to lift your IVR to the next level.
The authors wish to thank Karunesh Ahuja and Aditi Ranasaria for their contributions to this blog post.
1 Interactive Voice Response (IVR) Systems: Global Strategic Business Report, Research and Markets, January 2023.