In deciding where to automate, organizations are prioritizing internal efficiency goals over CX data about what their customers want, a CallMiner survey found.
Dive Brief
- More than two-thirds of organizations are underutilizing CX insights to guide their automation efforts, according to a CallMiner survey of 700 senior contact center and CX decision-makers released on Tuesday.
- While 71% said customer insights inform their decisions about what processes to automate, 76% said that their automation is driven by internal goals rather than customer needs. Respondents listed prioritizing reducing cost, increasing efficiency and reducing capacity with managing high-volume interactions as their top internal goals.
- About one-quarter of respondents described their automated customer experiences as very positive, and half described their automation efforts as mostly positive. Another one-quarter described their CX automation results as mixed or negative.
Dive Insight
Organizations are prioritizing internal efficiency goals and underutilizing their CX data about what their customers want in deciding aspects of their experience to automate.
Less than one-third of organizations are using CX insights derived from data sources like transcripts, operational signals and automation performance data to improve automated customer interactions. From these data sources, organizations can identify customer needs, behaviors and pain points.
“I think we need to start looking at customer friction as an opportunity signal,” Julie Geller, principal research director at Info-Tech Research Group, told CX Dive in an email. “Disconnected data, poor handoffs, conflicting policies or incentives that work against one another are prime places to tackle as your next AI opportunity.”
Many AI-powered tools are sold as ways to automate easy-to-answer customer inquiries that make up the bulk of interactions, everything from order status requests to password resets and billing inquiries or requests to upgrade a device or service plan. However, customers may have questions about these processes that might break the script or for which a generative AI-powered chatbot’s response is inadequate. Customers then need a human agent.
Notably, organizations with “very positive” automated customer experiences were more likely to escalate to human agents and draw on multiple CX data sources from during and after the actual interaction
That respondents in the “very positive” group escalate more frequently to human agents may not mean that their automation failed. Instead, it could suggest that they are more deliberate in determining when customers should transition from automation to human expertise
“There is a real risk of over-automation. When customers cannot reach a person, explain an unusual situation, or even challenge a decision, a small service issue can quickly become a trust issue,” Geller said. “Organizations need clear escalation points, strong data governance for accuracy and bias, and monitoring of customer outcomes, not just low-hanging-fruit metrics like productivity or cost savings.”
Leveraging automation performance data from chatbots, voicebots and interactive voice response systems, as well as omnichannel conversation transcripts can help organizations identify where customers drop off or escalate. Those insights can then be used to improve the self-service experience.
“Companies should start by defining the decision AI is intended to support and, most importantly, the customer outcome they are trying to improve,” Geller said. “It sounds simple, but it prevents organizations from executing a plug-and-play habit.”
