The Corporate U-turn on AI

The corporate U-turn on AI: your latest corporate governance news analysis examines how some firms tried to do too much, too fast. Now, they’re backtracking.

Artificial intelligence is a transformative technology; it’s astounding how rapidly it has become a part of daily business life, from every level up to the boardroom table. 

The automation is next-level, and the data processing is game-changing. The wealth of potential has meant many corporate leaders have raced to embrace it, to become poster companies for forward-thinking strategy. Clearly, however, some have gone too far too fast, speeding down a path that is rarely a straight line and hitting the rails along the way. 

There is now a growing number of high-profile cases revealing the problems that firms have encountered in their quest for AI adoption. You might consider them teething issues or maybe something more serious, but it’s crucial to analyse these now as a blueprint for further thinking about the pace of AI adoption that works for your company.

While many AI projects are succeeding, some corporations – particularly those that chose rapid adoption, filling many new roles with AI – have been forced to roll back or abandon their plans. In short, there’s an “AI remorse” occurring, where initial promises and optimism have been dealt something of a reality check as expected results didn’t materialise, and criticism mounted.  

Examples include:

• Klarna: The Swedish fintech company, which once championed its AI customer service as being as effective as 700 human agents, has reversed course and is now hiring human workers again after extensive feedback led to admissions that the AI element simply wasn’t living up to the human equivalent. CEO Sebastian Siemiatkowski admitted that focusing on cost-cutting over quality resulted in “lower quality” service from the AI chatbot.
• McDonald’s: As AI became a global buzzword in the early 2020s, McDonald’s rolled out an AI-based order system at 100 of its outlets in the United States. In 2024, the fast food giant suddenly announced that it would stop this project entirely. No official reason was given, but multiple reports had surfaced of the AI systems repeatedly getting orders wrong, fuelling customer frustration.
• Commonwealth Bank of Australia, one of Oceania’s biggest lenders, announced a major reversal of a decision to replace 45 human customer service employees with an AI chatbot. Once the chatbot was implemented, human employees said that their level of work spiked. In a statement, Commonwealth Bank said it “did not adequately consider all relevant business considerations” when announcing the redundancies.

That quote you just read – “not adequately considering all relevant business considerations” – strikes at the heart of this particular issue: companies realising that their big bold move to AI didn’t carry all the positives they thought it would. 

Indeed, a 2025 survey by technology research firm Orgvue found that of the business leaders who replaced employees with AI, 55% regretted the decision and 42% of generative AI pilot projects were abandoned in 2025, a substantial increase from the previous year.

This is an expected phenomenon – a symptom of when ambition outpaces thoughtful governance. This behaviour mirrors past corporate overcorrections, such as the mass hirings seen in the technology sector during the COVID-19 pandemic as the world moved online, which later resulted in major layoffs when companies realised their recruitment strategies didn’t make long-term sense. 

From a corporate leadership perspective, there are two reasons why we would see this coming:

• Any good corporate leader knows that part of their job is to be proactive and think outside the box. Often, this can manifest as aggressive positivity toward an innovation. In the rush, vital governance questions can be missed. 
• The media has marvelled at the capabilities of AI, as have industry experts and tech enthusiasts. From every major boardroom meeting to every weekly webinar, you’ll hear the narrative that AI is an overhaul that’s here to stay. In that kind of environment, corporate leaders will always be encouraged to act fast without double-checking every detail. The lack of any past analogues will also factor in. 

These factors give us bad governance decisions made with the best governance intentions. Decisions that made sense two years ago haven’t panned out. 

While there’s no denying that the adoption of AI is a tricky thing to get right from a board’s perspective, it does underscore the need to call a halt to rapid adoption if vital questions haven’t been answered yet. If you’ve done a thorough analysis, excellent. But if you haven’t, this might be the time to do so before company performance forces you into a retreat.

AI is the topic of this article, but it’s certainly not the problem. As many have said, it will undoubtedly remain a permanent and transformative feature of corporate life. Early missteps can happen, and they are avoidable. 

The real trick is balancing the need for innovation with proper fiduciary duty. For AI adoption, that means asking the tough questions early, ensuring that you can back up your decision-making with in-depth analysis from expert thinkers.   

Ultimately, the future of AI in the corporation requires a governance mindset that views AI as a powerful tool requiring careful, responsible stewardship. Showing it that respect takes time. If you don’t make that time available, you’re venturing into “magic-bullet” territory, expecting all the perks and planning for none of the drawbacks.