AI Boom in Customer Service Failing to Deliver Expected Results, New Research Finds

By Francisca Anuforo,

As businesses worldwide accelerate investments in artificial intelligence (AI) for customer service, a new study suggests that many organisations may be solving the wrong problems.

The Beyond the Chatbot: CX AI Report 2026, released by customer service AI company Charmedly, reveals that while AI adoption has become mainstream in customer experience (CX) operations, improvements in customer satisfaction have lagged behind expectations.

Based on a survey of 101 customer service professionals across seven industries and five regions, the report offers a rare frontline perspective on how AI is being used—and where it is falling short. Notably, 71 per cent of respondents were from Africa and the Middle East, regions that remain underrepresented in most global AI and CX research.

According to the findings, 69 per cent of organisations have deployed AI tools for customer service, while 62 per cent have implemented customer-facing chatbots. Yet despite these investments, satisfaction scores have not moved in the direction many businesses anticipated.

The report argues that the problem is not resistance to AI. In fact, support for AI-powered assistance was widespread among respondents. Instead, the issue lies in how many organisations are deploying the technology.

Researchers found that chatbot implementations are often designed primarily to reduce operational costs and deflect customer inquiries rather than improve issue resolution. As a result, many customers perceive chatbots as barriers rather than enablers of better service.

“The chatbot brief was wrong,” the report notes, arguing that organisations have prioritised efficiency metrics over customer outcomes.

One of the most significant findings is that information management—not AI capability—has emerged as the biggest productivity challenge facing customer service teams.

More than half of respondents identified faster access to accurate information as their top efficiency need, making instant knowledge retrieval the most requested AI feature. By contrast, real-time coaching, a feature heavily promoted by many AI vendors, ranked among the least desired capabilities.

The study also highlights a growing phenomenon often referred to as “shadow AI.” About 61 per cent of respondents reported using ChatGPT independently in their work without formal company guidelines, approved knowledge sources or governance frameworks.

Despite this widespread adoption, trust in AI-generated content remains limited. The report found that 82 per cent of customer service professionals routinely edit AI-generated responses before sharing them with customers, while only six per cent use AI outputs without modification.

Perhaps the most unexpected finding, however, was the emergence of emotional labour as a critical but largely invisible dimension of customer service work.

Nearly half of respondents who provided open-ended feedback voluntarily raised concerns about emotional strain, burnout and the psychological demands of handling difficult customer interactions. Significantly, these concerns surfaced without any direct questions on the topic.

According to Charmedly Founder, Toluwanimi Onakoya, this may be the most important insight from the research.

“We surveyed 101 CX professionals and found that 27 of them raised emotional labour without being asked. It was not on the list. That is not a footnote—it is the most important finding in the report,” he said.

The report suggests that as AI systems absorb routine customer interactions, human agents are increasingly left to manage complex, emotionally charged and escalated cases. Yet most AI performance metrics remain focused on speed, automation rates and cost reduction, leaving emotional workload largely unmeasured.

For organisations pursuing AI-driven transformation, the findings offer an important lesson: successful AI adoption may depend less on deploying more bots and more on empowering frontline workers with the right information, governance frameworks and support systems.

As enterprises continue investing billions in generative AI and automation technologies, the report raises a timely question: are companies optimising customer service for efficiency, or for better customer outcomes?

The answer, according to frontline workers, may determine whether the next wave of AI investment delivers meaningful value or simply more sophisticated frustration.