AI can perform a task without eliminating the job. And sometimes, it creates more work for the people who remain.
Companies can spend months planning an AI rollout, testing its capabilities and anticipating how jobs will change. Then it reaches production, and the work doesn't always unfold the way anyone expected. That's when staffing decisions that seemed reasonable on paper can become much harder to justify.
When the staffing math doesn't work out
At Syndio, a company that develops pay intelligence software, a shift toward an AI-centered business model led to staffing decisions its CEO later acknowledged getting wrong.
In May, Syndio eliminated several positions, including that of labor economist Jonathan Vidales. Around the same time, the company redesigned other jobs, adding responsibilities and sales quotas for employees who remained.
CEO Maria Colacurcio expected those employees would welcome the opportunity, particularly with the job market becoming less favorable to workers. "We assumed they wouldn't blink," she wrote in an August account.
Instead, several resigned. By August, Syndio had rehired Vidales into a newly opened position. Colacurcio acknowledged that the company had underestimated the value of its existing employees and their ability to adapt as the business changed.
This wasn't simply a case of an AI system failing to replace a worker. Syndio was changing its business model, and its staffing assumptions proved more complicated than anticipated.
Commonwealth Bank of Australia offers another example, although its experience was different.
After introducing an AI-powered voice bot, the bank announced plans in July 2025 to eliminate 45 customer-service positions. But employees and their union reported that call volumes were increasing, with the bank offering overtime and assigning team leaders to answer calls.
Weeks later, Commonwealth reversed the decision and apologized, acknowledging that it had not adequately considered the business requirements when assessing the positions.
Employees were given the option to remain in their roles, seek redeployment or leave. The reversal gave affected employees an opportunity to stay, although the uncertainty had already taken a toll on workers.
Neither experience establishes that AI itself failed. Both raise questions about how employers assess the consequences of introducing it.
Understanding the work before changing the workforce
Research released in October from talent assessment firm SHL suggests employers may have an incomplete understanding of the people already working for them.
Of 2,036 U.S. and U.K. workers surveyed by Propeller Insights, only 42% believed their managers fully understood their skills beyond their current positions. Another 84% wanted employers to identify internal opportunities matching their capabilities.
"Many organizations are treating AI as a way to take cost out of existing work rather than as a reason to redesign it, and that is the risk."
Lucy Beaumont
senior vice president, SHL
Those findings raise an important question: How confidently can a company redesign or eliminate a position if it doesn't fully understand the skills and responsibilities involved?
SHL also reported that 19% of U.S. respondents said they had personally been laid off, replaced by AI and subsequently asked to return by the same employer. However, the research did not verify that AI actually caused those layoffs, SHL senior vice president Lucy Beaumont told TechTarget.
"Many organizations are treating AI as a way to take cost out of existing work rather than as a reason to redesign it, and that is the risk," Beaumont said.
Gartner's research adds another perspective. In a survey of 321 customer-service and support leaders, 85% reported expanding human agents' responsibilities as AI changed their work. At the same time, 31% had implemented or planned AI-driven frontline layoffs through the first quarter of 2027.
Those findings aren't necessarily contradictory. AI can reduce some responsibilities while expanding others. The challenge is understanding how those changes fit together before making permanent staffing decisions.
A successful pilot isn't the end of the assessment
Assessing the operational and workforce impact of new technology is already an established part of sound IT management. AI shouldn't be an exception.
But even a successful pilot cannot establish everything that will happen in production.
"Redesign work around AI, not just around tools. Decide which tasks AI should take on and where human judgment, context and relationships add the most value."
Lucy Beaumont
senior vice president, SHL
A system might answer customer questions accurately during testing without showing how the volume of more complicated calls will change. An AI assistant might process routine requests faster while leaving employees with more exceptions, additional oversight or greater responsibility for resolving problems.
And when employees are reassigned or positions eliminated, the remaining work doesn't necessarily disappear with them.
"Redesign work around AI, not just around tools. Decide which tasks AI should take on and where human judgment, context and relationships add the most value," Beaumont explained.
That doesn't mean companies should avoid planning staffing changes until an AI implementation is complete. They need to understand the possible outcomes, including changes to job descriptions, responsibilities and staffing needs.
But projected productivity gains should not automatically become permanent workforce reductions.
The more consequential test comes when AI is operating within the business, alongside employees, existing systems and the complications that may not have appeared during development.
Some of those complications will only become apparent through experience. Even a thorough assessment cannot anticipate everything. Companies need enough flexibility to learn from what happens and change course when necessary.
What AI testing can show — and what CIOs still need to know
- AI successfully handles certain tasks
What happens to the complete workflow?
- Productivity improves during testing
Will those gains persist in production?
- AI reduces time spent on particular activities
What additional work falls on remaining employees?
- New roles have been designed
Do employees have the skills, capacity and willingness to perform them?
- Financial projections show potential savings
Do actual savings justify permanent staffing changes?
The people doing the work need a voice
One way to improve those assessments is to involve employees who understand how the work actually gets done.
Not every employee needs to participate in an AI project. But a representative group can help identify exceptions, informal responsibilities, dependencies and tasks that don't necessarily appear in job descriptions or process documentation.
That knowledge also matters when redesigning jobs. Employees can help determine where automation will genuinely save time, where new responsibilities will emerge and what training or development people will need as their work changes.
Their participation isn't simply a matter of treating people well, although that matters. It can lead to better technology implementations and a better understanding of the workforce the business will need afterward.
Beaumont also sees a need for technology and HR leaders to work together on those decisions.
"In my experience, the poorest decisions come when AI programs are led by technology alone," she said, explaining that questions about employees' skills, changing roles and adoption can become afterthoughts.
Of course, getting candid participation may be difficult when employees suspect they're helping introduce technology that could eventually eliminate their jobs.
And those concerns are hardly surprising. A September Gallup survey found that 27% of U.S. workers worried technology could make their jobs obsolete, up from 13% in 2017.
Employers can't solve that problem simply by promising AI will supplement rather than replace people, particularly if they don't yet know what the eventual staffing impact will be.
Honesty matters, but so does giving employee feedback real weight. Commonwealth Bank's reversal followed challenges from workers and their union over what was happening in its call centers. Employees' operational experience provided a perspective that management's initial assessment had not adequately reflected.
CIOs and business leaders need ways to gather that information during planning, implementation and production — and to act on it when the evidence challenges their assumptions.
Before anticipated savings become actual cuts
Some AI implementations will genuinely reduce the number of people needed to perform particular work. Others will change jobs, create new responsibilities or free employees to take on work the business previously lacked the capacity to do.
There is no reason to assume every existing position must remain unchanged.
But there is also no reason to assume that the work AI can perform tells the whole story about which employees a company will need.
Before turning projected efficiencies into permanent staffing reductions, CIOs, HR, finance and business leaders need to understand what the organization has actually gained, what work remains and what additional costs or responsibilities the technology has created.
That includes evaluating whether employees can be retrained or reassigned as the work changes.
An organization can revise an AI implementation, change a workflow or adjust its operating model. Rebuilding lost expertise, replacing experienced employees and restoring trust could prove considerably harder.
The challenge isn't simply knowing what AI can do. It's understanding what the business and its people will need once AI is doing it.
James Alan Miller is a veteran technology editor and writer and Lead Editor for CIO News at TechTarget. He directs coverage of enterprise technology strategy, AI, software, data, infrastructure and the decisions shaping how CIOs manage increasingly complex IT environments.
