The artificial intelligence boom promised businesses faster operations, lower costs and widespread automation. However, some companies are now reversing course after discovering that AI systems often come with hidden expenses, reliability issues and limitations in handling complex tasks.
Several businesses that reduced human staff to replace roles with AI are now rehiring employees to improve accuracy, customer satisfaction and decision-making.
Around 32% of companies that downsized workers during AI adoption are reportedly rehiring for those same positions, according to industry data. Business leaders have cited problems including poor customer experiences, AI errors and the loss of valuable institutional knowledge.
The shift highlights a growing realisation among companies that artificial intelligence works best as a support tool rather than a complete replacement for human expertise.
Major Companies Reverse AI-Only Strategies
Several well-known companies have faced challenges after rapidly adopting AI-based systems.
Klarna, the Swedish fintech company, previously reduced its customer service workforce after introducing AI tools. The company replaced hundreds of customer support tasks with artificial intelligence but later faced criticism over unresolved customer complaints and difficulties handling complicated cases.
Read More:Â Meet the AI Drone That Hunts and Kills Mosquitoes Mid-Air
Following these issues, Klarna began increasing human involvement in customer service operations.
Commonwealth Bank of Australia (CBA) also experienced problems after replacing 45 customer service workers with an AI voice bot. The system struggled with customer interactions, leading to unresolved calls and negative feedback. The bank later apologised and brought back human staff to improve service quality.
IBM reduced parts of its human resources operations through automation. While AI successfully managed routine requests, the company found that sensitive employee issues, ethical concerns and complex decisions required human judgment.
Technology companies including Google, Meta and Salesforce have also increased hiring for roles focused on managing, monitoring and improving their generative AI systems after earlier workforce reductions.
Meanwhile, McDonald’s paused its AI-powered drive-thru ordering experiment after customers shared videos online showing incorrect orders. The fast-food chain eventually ended the pilot and returned to traditional human-supported operations.
Ford Says AI Cannot Replace Experienced Engineers
The automobile industry has also seen limits in AI adoption.
Ford Motor has acknowledged that artificial intelligence and automation alone cannot replace experienced engineers when designing and building high-quality vehicles.
The company has strengthened its engineering teams by hiring, rehiring and promoting around 350 experienced technical specialists. Ford said these experts are helping improve vehicle quality, mentor younger engineers and refine AI-powered quality systems.
The move follows years of investment in automation and artificial intelligence aimed at improving manufacturing efficiency, identifying defects and speeding up vehicle development.
Ford executives believe AI can support engineers but cannot replace years of practical experience and problem-solving skills developed through building real vehicles.
Human Expertise Remains Critical in AI Era
The recent shift does not signal an end to artificial intelligence adoption. Instead, companies are changing their approach by combining AI tools with human oversight.
Businesses increasingly recognise that AI performs well in repetitive tasks, data processing and automation. However, human workers remain essential for creativity, judgment, customer relationships and handling unexpected situations.
The experience of companies such as Klarna, CBA, IBM and Ford suggests that the future workplace may not be humans versus AI, but humans working alongside AI.
As businesses continue investing billions into artificial intelligence, many are discovering that replacing people entirely may not deliver the efficiency and quality they originally expected.
