Artificial intelligence represents a breakthrough in handling routine, low-complexity customer support tasks. However, some companies are so focused on rapid growth that they attempt to automate nearly everything.
Yet real-world examples show that excessive automation can alienate customers, leading to more negative reviews and unfavorable media coverage.
This article explores the risks of fully automated support, explains where AI works best and where human support remains critical, and shows how both approaches can be combined into a more efficient system.
You’ll also learn how the winners of PissedConsumer’s Consumers Choice Award combine technology and human interaction in their support practices, and why personal connection strengthens customer loyalty to a brand.
- Why AI customer support savings may not last
- The illusion of “fully automated customer support”
- Where AI actually works extremely well
- Where humans are irreplaceable
- Why the best support teams combine AI and humans
- Final thoughts
Why AI customer support savings may not last
To cut customer support costs, some companies are rushing to implement artificial intelligence. But many of them fail to realize that customers want not just faster support, but support that better understands their needs.
According to Gartner forecasts, by 2027, 50% of companies that cut staff due to AI implementation will rehire employees to perform similar functions.
So far, however, relatively few companies have made such cuts. A Gartner survey conducted in October 2025 showed that only 20% of executives had reduced their agent workforce due to AI. Meanwhile, 55% of companies reported implementing AI without changing staffing levels – AI now helps agents handle more customer inquiries but doesn’t replace them.
According to the research firm, AI isn’t yet mature enough to fully replace the experience, empathy, and judgment provided by human agents. Therefore, relying solely on AI at this stage is premature and could lead to unpredictable consequences. Furthermore, ever since automated response systems appeared in call centers, customers have often complained about the difficulty of reaching a real person. They are even less willing to adapt to limited-capability AI agents, believing that a human will better understand their problem.
As a result, Gartner predicts that half of the organizations expecting significant staff reductions due to AI will abandon these plans, as the idea of “agentless” service will prove unfeasible.
The illusion of “fully automated customer support”
Replacing humans with AI in customer support is already having negative consequences.
Klarna is now rehiring employees after replacing 700 workers with artificial intelligence following problems with its customer support service. Once one of the most vocal advocates of AI, the company now acknowledges that its automation tools failed to provide empathy and a deep understanding of consumers’ problems. Customers complained that the AI responses were generic, repetitive, or unable to solve real problems.
Here’s another example of how unwanted automation can frustrate customers.
After Hilton introduced virtual agents to assist with bookings in 2026, social media users began reporting that the AI agents couldn’t answer even simple questions, such as whether the hotel’s outdoor pool was open. Moreover, customers claimed that the AI voice assistants presented themselves as live agents and repeatedly avoided requests to speak with a real human.
Companies striving to automate as much as possible need to ask themselves: are these changes in customer service actually what consumers want?
Even the most sophisticated generative AI systems may fail to fully understand the context of a question, and deviations from pre-programmed scenarios can result in irrelevant responses. Moreover, human interaction involves emotional nuance and interpersonal connection that machines are unlikely to replicate, regardless of how advanced or natural their responses appear.
Meanwhile, AI delivers the most value behind the scenes by helping companies optimize workflows, collect and analyze data, and categorize customer support requests.
Where AI actually works extremely well
AI can take over some of the routine tasks performed by human agents, freeing them up to handle more complex inquiries. Examples of tasks that can be delegated to artificial intelligence include:
- handling routine customer inquiries such as order status requests, shipping updates, refund policies, and basic troubleshooting questions;
- finding relevant knowledge base answers;
- categorizing and routing tickets;
- detecting sentiment;
- suggesting actions based on historical customer data;
- gathering data from multiple channels to provide a complete picture of customer interactions.
Where humans are irreplaceable
Consumers have grown accustomed to some level of automation in customer support interactions. However, the need for human empathy hasn’t disappeared. Here are some situations where the human touch remains invaluable:
- dealing with strong emotions and adapting to the customer’s communication style and pace;
- resolving complex issues that require critical thinking and negotiation skills;
- managing unusual cases that may require decisions beyond standard policies;
- handling sensitive situations where customer trust is at risk.
In these cases, empathetic human support becomes a critical contributor to customer loyalty and retention.
Why the best support teams combine AI and humans
Companies that view AI initiatives solely through the lens of workforce reduction risk damaging customer relationships. They focus on minimizing the cost per ticket above all else, but, as the example of Klarna shows, issue resolution and long-term brand perception often take a back seat in this approach.
As a result, customers become frustrated when they receive irrelevant responses and cannot reach a real person. They feel that no one at the company is taking responsibility for solving their problem. When customers hit a dead end with an AI bot, their frustration moves to public review platforms. PissedConsumer data consistently shows that negative reviews spike when customers feel blocked from reaching a live agent.
Ultimately, the best support teams succeed not because they’ve automated more processes, but because they’ve created a model in which AI and humans perform the tasks they’re best suited for.
AI handles:
- triage;
- summarization;
- response drafts;
- predictive analytics.
Humans handle:
- final judgment;
- tone adjustment;
- sensitive customer interactions;
- escalation management;
- relationship repair.
This approach combines the speed and analytical capabilities of AI with the empathy and critical thinking that humans excel at.
Final thoughts
Fully automated support may reduce costs in the short term, but the hidden costs associated with dissatisfied customers and declining loyalty can eventually outweigh those savings. In moments of frustration, confusion, or urgency, customers still seek human understanding – something automation cannot fully replace.
The companies most likely to succeed in the long term are those that use AI to empower their teams while preserving the empathy, accountability, and critical thinking that only humans can provide.
Want to see how top brands achieve 5-star customer satisfaction? Download the full Consumers Choice Award 2026 Report to discover the metrics behind the winners.
Legal disclaimers:
- While every effort has been made to ensure the accuracy of this publication, it is not intended to provide any legal, medical, accounting, investment or any other professional advice as individual cases may vary and should be discussed with a corresponding expert and/or an attorney.
- All or some image copyright belongs to the original owner(s). No copyright infringement intended.
