Quick Buzz Feed
AI assistance is only as good as the data it’s grounded on; decisions based on any single data point or disjointed systems can create unintended friction points.
A recent TheyDo survey of 1,000 business decision-makers revealed that nearly 2 in 5 leaders experienced unintended negative consequences for customer experience (CX) due to using AI-powered tools in their decision-making. Among those facing CX issues, roughly half reported an increase in customer complaints or support queries, and a similar number observed confusing or inconsistent customer experiences. TheyDo's research clarifies that AI itself isn't necessarily the cause of these problems, but rather highlights a pattern where AI assistance, while speeding up decisions, might rely on data that is fragmented or inconsistent across various systems and teams. As Jochem van der Veer, co-founder and CEO at TheyDo, explained to CX Dive, a customer's journey often spans more boundaries than any single system or department can track, leading decisions that seem logical in one area to cause friction elsewhere.
The challenge arises because customer journey data is scattered across numerous back-end systems, making decisions based on limited or disjointed data points prone to creating unintended friction. Julie Geller, principal research director at Info-Tech Research Group, emphasized that if the foundational data is incomplete, outdated, or biased, any AI recommendations built upon it will similarly be flawed. TheyDo’s research suggests that the crucial element is generating comprehensive context from these disparate data points, which AI-powered tools can facilitate through robust integrations that ensure data remains current and accurate. However, AI-generated CX recommendations alone may not be sufficient, as TheyDo found that CX leaders were 3.5 times more likely to trust AI's output when they could visualize the underlying customer journey that informed the recommendation. Geller highlighted that the true value of AI lies not in making decisions, but in helping leaders quickly identify issues, understand their root causes, and determine the most impactful interventions. The survey also indicated that nearly one-third of CX problems first surface during customer service interactions. Van der Veer pointed out that service teams often detect the downstream consequences of upstream decisions before the rest of the organization, underscoring the necessity for these signals to inform decisions made by product, marketing, and operations teams. Despite the benefits of AI-powered summaries and insights into customer feedback, CX leaders must not abandon direct engagement—speaking with, listening to, and observing customers. Jon Picoult, founder and principal of Watermark Consulting, noted that AI engines are not flawless in distilling key themes from all data, and skilled CX practitioners bring a unique ability to conduct qualitative research, probing customers to gain a deeper understanding of their nuanced needs, desires, hopes, and fears, an insight that AI tools cannot fully replicate.