Employers rethink their AI bet: half of companies need people again

Companies that rushed to replace staff with AI bots are encountering a reverse effect: automation hasn't closed all tasks, and in several cases work quality and speed have declined. Against this backdrop, around half of such employers may return to hiring people as early as next year.

Where the Pullback Came From

The market increasingly discusses a phenomenon already being called "AI-washing." This refers to cases where layoffs are explained by artificial intelligence implementation, though the real reasons may be far more mundane: cost pressures, weak demand, restructuring, or attempts to quickly improve metrics for investors. In such scenarios, AI becomes not so much a working tool as a convenient justification for tough personnel decisions.

The problem is that glossy presentations about complete automation poorly match the actual work of companies. Management finds it easy to announce that a chatbot, email generator, or AI-powered support system will replace an entire function. It's far harder to admit that much of office and customer work consists of exceptions, clarifications, responsibility, and constant task switching. It's precisely in these areas where the pullback from overly aggressive automation begins.

Where Bots Let Them Down

In practice, AI performs best with templated actions: drafting a response, categorizing data, assembling summaries, finding the right fragment in documents. But when a process goes beyond a pre-defined scenario, efficiency drops sharply. People have to fix errors, and the promised savings start to erode due to repeated checks, customer complaints, and lost time across teams.

Problems most often emerge in these areas:

• non-standard customer inquiries where context and tone matter
• fact-checking, numbers, and legally significant wording
• coordination between departments where decisions depend on internal company history
• processes where errors lead to direct financial or reputational losses

Another unpleasant effect is the loss of internal knowledge. When people leave, so does the experience of those who understood product nuances, customers, and internal operations. Later it turns out that bots can respond quickly but can't take responsibility for controversial decisions. So business doesn't just need to reopen positions but rebuild competencies that were too hastily written off as unnecessary.

An additional problem involves hidden maintenance costs. After cutbacks, companies still spend resources on model tuning, answer monitoring, process rework, and error analysis. Part of the saved payroll budget ends up being consumed by new operational expenses, and teams begin working in a mode of constant manual bot supervision.

Why Hiring Is Returning

The return to hiring doesn't mean AI has failed as a technology. Rather, the market is starting to sober up after a period of inflated expectations. Companies see that neural networks are useful as an acceleration layer on top of a team, not as a universal replacement for employees. Where AI was implemented as a helper, results are usually better: people do less routine work, prepare materials faster, and process more tasks without quality drops.

A more realistic model is now forming: leave repetitive operations to automation, and humans handle decision-making, complex customer work, quality control, and accountability for outcomes. For many employers this means a new hiring round, but for different roles. They don't need just executors but employees who can work alongside AI, verify its conclusions, and quickly intervene when automation fails.

What This Means

The main takeaway is simple: it turned out easier to lay people off under the banner of "AI replaced us" than to build a sustainable process without them. Business appears to be moving from a fashionable gesture to a practical approach where neural networks truly reduce routine but don't eliminate the value of human experience, context, and responsibility.