New Research Reveals Why Replacing Workers With AI Is Failing

For the past two years, artificial intelligence has been sold as the biggest workplace revolution since the internet. Executives promised leaner operations, lower costs, and a future where software could perform the work of thousands of employees. Across Silicon Valley and beyond, companies responded by trimming their payrolls while pouring billions into AI.

Now, a growing body of evidence suggests that many of those decisions may have been made far too quickly.

A major new study from research and advisory firm Gartner has found that replacing employees with AI is failing to produce the financial gains many business leaders expected. Instead of creating a clear competitive advantage, companies that cut jobs to fund AI initiatives are often performing no better than businesses that kept their workforces intact.

Companies Are Cutting Jobs Before Seeing AI Deliver Results

The excitement surrounding artificial intelligence has reshaped boardroom priorities at a remarkable pace. Businesses have raced to deploy chatbots, automation platforms, coding assistants, and generative AI systems in hopes of reducing labor costs while increasing productivity.

Gartner wanted to determine whether that strategy was actually working.

The firm surveyed 350 global business executives whose organizations each generate at least $1 billion in annual revenue. The results paint a much different picture than many headlines have suggested.

According to the research, 80 percent of executives whose organizations had piloted AI or autonomous technologies reported reducing their workforce. However, those layoffs were often carried out before companies had evidence that AI would generate meaningful returns.

Perhaps the most surprising finding was that businesses which reduced headcount did not outperform companies that retained their employees. Financial gains remained largely similar across both groups, suggesting that replacing workers with AI was not creating the advantage executives hoped to achieve.

Helen Poitevin, a Gartner vice president analyst who helped lead the research, believes many organizations have been looking for value in the wrong place.

“Looking only at layoffs is shortsighted in terms of getting value from AI,” she told Fortune. “Chasing value only through headcount reduction is likely to lead most organizations down a path of limited returns.”

That conclusion challenges one of the most common assumptions surrounding AI adoption. Many executives believed reducing payroll would naturally improve profitability once software took over routine work. Instead, the survey indicates that eliminating experienced employees often removes valuable knowledge while doing little to improve overall business performance.

The AI Boom Created a Race Few Wanted to Lose

The findings become easier to understand when viewed alongside the enormous pressure companies have faced over the past two years.

Artificial intelligence quickly evolved from an emerging technology into a competitive necessity. Every major earnings call seemed to include AI announcements. Investors rewarded companies that embraced automation, while organizations that appeared hesitant risked looking outdated.

Rather than asking whether AI was mature enough to replace workers, many executives began asking how quickly they could implement it.

This competitive environment created what many analysts describe as a fear of being left behind.

Businesses rushed to deploy AI across customer support, marketing, software development, finance, and administrative operations. In many cases, layoffs became the easiest way to free up funding for expensive AI infrastructure and software investments.

Several high-profile technology companies reduced thousands of positions while simultaneously increasing spending on AI development. Those announcements fueled public concerns that artificial intelligence would rapidly replace millions of white-collar jobs.

Yet Gartner’s findings suggest that those workforce reductions often had little connection to measurable AI success.

Instead, many organizations appear to have treated layoffs as part of an experiment rather than the outcome of proven business improvements.

Poitevin believes the workforce reductions represent isolated attempts to test AI rather than permanent restructuring.

“It seems to us to be a kind of one-time exercise by many in small amounts, but not what translates to getting full ROI from their AI investment,” she explained.

That distinction matters because it suggests many companies are still searching for the right balance between automation and human expertise instead of confidently moving toward fully automated workplaces.

Businesses That Keep Their Employees Are Seeing Better Results

One of the clearest patterns to emerge from Gartner’s research is that AI performs best when it supports employees instead of replacing them.

Companies reporting the strongest returns on investment were generally using AI as a tool for what Gartner describes as “people amplification.”

Rather than eliminating positions, these organizations equipped workers with AI systems that helped them complete repetitive tasks faster, organize information more efficiently, and spend more time solving complex problems.

The technology became an assistant rather than a substitute.

This approach recognizes something many businesses are beginning to rediscover. Most jobs consist of dozens of different responsibilities. AI may excel at handling certain repetitive tasks, but it often struggles with judgment, relationship building, creativity, negotiation, and unexpected situations that require human experience.

A customer service representative, for example, does much more than answer routine questions.

They calm frustrated customers, recognize emotional cues, adapt conversations based on context, identify unusual situations that fall outside company policies, and preserve relationships that software can easily damage.

Removing that human element entirely often creates new problems instead of solving existing ones.

Businesses that understand these limitations appear to be generating greater value from AI because they are improving human productivity instead of attempting to eliminate human involvement altogether.

Why AI Still Falls Short of Replacing Human Judgment

The excitement surrounding artificial intelligence has sometimes created the impression that computers are rapidly approaching human-level thinking across every profession.

Reality remains much more complicated.

Today’s AI systems are exceptionally good at recognizing patterns, summarizing information, generating text, writing software code, and automating repetitive processes. Those capabilities have already transformed many workplaces.

However, AI also continues to produce incorrect information with remarkable confidence.

Large language models can fabricate facts, misunderstand context, misinterpret customer intentions, or generate responses that sound convincing while being entirely inaccurate.

These errors, often called hallucinations, require ongoing human oversight.

That oversight introduces costs that many early business forecasts underestimated.

Organizations still need employees to review AI output, verify accuracy, monitor security risks, ensure regulatory compliance, and intervene whenever automated systems encounter situations they cannot properly handle.

In industries involving healthcare, finance, law, engineering, cybersecurity, or customer support, those responsibilities remain critical.

Replacing experienced professionals too quickly can create risks that outweigh the savings generated through workforce reductions.

Institutional knowledge represents another major challenge.

Employees accumulate years of experience understanding company history, customer expectations, internal processes, and relationships that cannot simply be transferred into an AI model.

Once those workers leave, much of that knowledge disappears with them.

Companies frequently discover that replacing experience is considerably harder than eliminating payroll expenses.

Earlier Research Was Already Pointing in the Same Direction

Gartner’s findings do not exist in isolation.

Over the past year, multiple independent studies have raised similar questions about whether AI investments are generating meaningful financial returns.

Researchers at MIT previously found that the overwhelming majority of organizations adopting AI were failing to produce measurable revenue growth from those investments.

Despite enormous spending across industries, relatively few companies reported meaningful business improvements directly linked to AI implementation.

Other industry surveys have also shown that many generative AI projects never move beyond the pilot stage.

Some organizations abandon deployments because of poor data quality.

Others encounter escalating implementation costs, security concerns, integration challenges, or uncertainty about how to measure success.

These obstacles rarely receive the same attention as dramatic announcements about AI replacing workers.

Yet they represent the everyday reality facing businesses attempting to integrate complex technology into existing operations.

For executives, the challenge is becoming less about whether AI has potential and more about identifying where it genuinely creates value.

That process is proving slower and more complicated than many early predictions suggested.

Real-World Examples Show the Limits of AI-Only Workforces

The conversation around AI replacing employees is no longer theoretical. Several companies that initially embraced aggressive automation have quietly adjusted their strategy after discovering that AI could not consistently deliver the results they expected.

One of the most widely discussed examples is Klarna.

The Swedish fintech company made headlines in 2024 after announcing that its AI assistant could perform the work of 700 customer service representatives. The announcement quickly became one of the most cited examples of AI replacing human workers, reinforcing the belief that large-scale workforce reductions were becoming inevitable.

Behind the scenes, however, the story evolved.

As customer service demands continued to grow, the company began rebuilding parts of its support operation. While AI remained an important tool, human agents proved essential for handling more complex conversations, resolving unusual cases, and maintaining customer satisfaction.

The shift highlighted an important lesson. Measuring success solely by how many employees can be replaced ignores the quality of service customers actually receive.

Gartner analysts believe this lesson is becoming increasingly common across industries.

In another Gartner survey focused specifically on customer service organizations, researchers found that only 20 percent of businesses had actually reduced customer support staffing because of AI. The overwhelming majority chose to maintain their workforce while integrating AI into daily operations.

Emily Potosky, Senior Director of Research for Gartner Customer Service & Support, explained why many organizations remain cautious.

“AI simply isn’t mature enough to fully replace the expertise, empathy, and judgment that human agents provide. Relying solely on AI right now is premature and could lead to unintended consequences.”

That observation reflects a growing understanding that customer interactions involve far more than answering straightforward questions.

Customers often contact companies when something has gone wrong. They may be frustrated, confused, or worried. Resolving those situations frequently requires emotional intelligence, flexibility, and judgment that current AI systems still struggle to replicate consistently.

Some Layoffs May Have Had Little to Do With AI

Another factor complicating the discussion is that not every AI-related layoff appears to have been caused by automation itself.

Over the past two years, numerous companies announced workforce reductions while simultaneously promoting major AI initiatives. That created a public perception that artificial intelligence was directly replacing employees.

The reality may be more nuanced.

Many technology companies have faced slower revenue growth, changing market conditions, and pressure from investors to reduce spending. Cutting staff became one way to free up capital for expensive AI infrastructure while also improving short-term financial results.

Some industry observers argue that AI has occasionally become a convenient explanation for decisions that businesses might have made regardless.

OpenAI CEO Sam Altman previously suggested that some organizations may be engaging in what has become known as “AI washing,” where layoffs are attributed to artificial intelligence even though other financial or operational reasons played a larger role.

“I don’t know what the exact percentage is, but there’s some AI washing where people are blaming AI for layoffs that they would otherwise do, and then there’s some real displacement by AI of different kinds of jobs,” Altman said.

That distinction matters because it changes how people interpret the broader employment picture.

Rather than signaling an unstoppable wave of automation, some layoffs may simply reflect ordinary business restructuring taking place during a period when AI dominates corporate messaging.

Why Human Skills Are Becoming More Valuable Instead of Less

Ironically, the rise of artificial intelligence may be increasing the value of uniquely human abilities rather than eliminating them.

AI performs exceptionally well when tasks follow predictable patterns. It can summarize documents, draft emails, analyze large datasets, generate software code, and answer routine questions within seconds.

Those strengths are undeniable.

The challenges begin when situations become unpredictable.

Negotiating with a difficult client, mentoring a junior employee, resolving workplace conflict, making ethical decisions, leading teams through uncertainty, or recognizing subtle changes in customer behavior remain areas where humans continue to outperform machines.

These capabilities rely on context, experience, emotional awareness, and intuition developed over years of interaction.

AI can assist with many of these activities, but it cannot reliably replace the people performing them.

As organizations gain more practical experience with AI, many are discovering that productivity comes from combining machine efficiency with human judgment rather than choosing one over the other.

That explains why companies reporting the highest returns are generally investing in employee training alongside AI deployment.

Workers who understand how to use AI effectively often complete repetitive tasks more quickly while dedicating more time to strategic thinking, creativity, and relationship building.

Instead of shrinking the workforce, many organizations are reshaping it.

New roles focused on AI oversight, prompt engineering, governance, quality assurance, cybersecurity, and workflow optimization are already emerging across industries.

The demand is shifting rather than disappearing.

History Suggests Technology Changes Jobs More Than It Eliminates Them

Concerns about automation replacing workers are hardly new.

During the Industrial Revolution, machines transformed manufacturing. Decades later, computers reshaped office work. The internet changed retail, publishing, entertainment, and communication.

Each technological breakthrough created anxiety about widespread unemployment.

Yet history shows that while technology eliminates certain tasks, it also creates entirely new industries and occupations.

Artificial intelligence appears to be following a similar pattern, although the speed of change is considerably faster.

Economists have pointed to a concept known as the Jevons paradox, which suggests that when technology makes something more efficient, overall demand often increases rather than decreases.

Apollo Global Management Chief Economist Torsten Slok has argued that this principle may apply to AI as well.

If businesses become dramatically more productive, they may eventually expand operations, launch new products, serve additional customers, and create different kinds of employment instead of permanently shrinking their workforce.

Even Anthropic CEO Dario Amodei, who previously warned that AI could eliminate large numbers of entry-level white-collar jobs, has recently acknowledged that the future may involve far more collaboration between humans and AI than initially expected.

That does not mean disruption will disappear.

Some occupations will inevitably change, and certain repetitive responsibilities will continue moving toward automation.

The larger question is whether entire professions vanish or simply evolve.

Current evidence increasingly favors the second outcome.

Companies May Soon Be Hiring Back the Workers They Let Go

Perhaps the most striking prediction comes from Gartner itself.

The research firm expects that by 2027, half of the companies that reduced customer service staff because of AI will hire people back to perform similar functions, even if those positions carry different job titles.

That forecast reflects an emerging realization across the business world.

Replacing experienced employees may reduce payroll costs in the short term, but rebuilding lost expertise often proves far more expensive.

Institutional knowledge cannot be downloaded overnight.

Relationships with customers take years to develop.

Employees understand company culture, historical decisions, internal processes, and countless unwritten practices that rarely appear in manuals or databases.

Once those people leave, organizations frequently discover that replacing them involves much more than filling an empty position.

It requires rebuilding experience from scratch.

For companies that moved too aggressively toward automation, rehiring may become the fastest path back to stable performance.

That possibility also offers reassurance for workers worried that AI has permanently closed the door on entire careers.

Many roles are likely to return in new forms that emphasize supervising AI systems, interpreting their outputs, and solving problems machines cannot.

The Companies Winning With AI Aren’t Trying to Replace Everyone

The biggest lesson from Gartner’s research is not that AI has failed.

Far from it.

Artificial intelligence continues to deliver meaningful improvements across software development, research, healthcare, logistics, finance, manufacturing, and countless other fields.

The technology is becoming more capable every month.

What appears to be failing is the belief that replacing people is the fastest route to profitability.

The organizations seeing the strongest returns are approaching AI differently.

They are treating it as a productivity tool rather than a workforce replacement strategy.

Employees use AI to reduce repetitive work, accelerate research, organize information, draft content, and automate routine processes while continuing to provide the judgment, creativity, communication, and leadership that businesses still depend on.

That combination is producing stronger financial outcomes than layoffs alone.

The early years of the AI boom have been dominated by dramatic headlines predicting massive job losses and fully automated workplaces.

The newest evidence tells a more balanced story.

Artificial intelligence is changing how people work, but the companies achieving the greatest success are discovering that technology performs best when it strengthens human capability instead of attempting to replace it.

As businesses move beyond the excitement of AI’s first wave, many are finding that their most valuable asset was never the software they purchased. It was the people they nearly replaced.