The AI boom has pushed companies to move quickly, sometimes faster than their own workers, customers, and internal systems can handle. From AI-driven layoffs to search engines forcing AI answers into results, the pressure to automate everything is creating a new divide between leaders who see AI as the next operating system for business and users who simply want products that still work.
A recent TechCrunch Equity discussion framed this tension through a blunt question: what happens when companies become too “AI-pilled”? The conversation pointed to a growing problem in the technology sector. Executives are not only adopting AI tools. In some cases, they are reorganizing companies, cutting jobs, changing products, and reshaping customer experiences around the assumption that AI can replace large parts of human work.
The concern is not that AI is useless. The concern is that some companies are treating it as a universal answer before they fully understand the jobs, workflows, and customer needs they are trying to automate.
The debate has become sharper because AI is no longer just a product feature or an investor pitch. It is now influencing headcount, budgets, user interfaces, and long-term business planning. TechCrunch’s video highlighted comments from Box founder Aaron Levie, who warned that the people deciding AI can replace workers may be the least familiar with what those workers actually do.
That warning lands at a sensitive time. Companies across the tech sector have spent the past two years telling investors that AI will make their teams more efficient. But efficiency has increasingly become tied to layoffs, restructurings, and aggressive automation plans. In one recent example cited by TechCrunch, ClickUp cut 22 percent of its workforce as it leaned further into AI agents.
The logic behind these moves is easy to understand from a boardroom view. If AI agents can handle support tasks, workflow coordination, coding assistance, content production, or internal operations, companies believe they can reduce costs and move faster. But that logic becomes risky when leaders underestimate the judgment, context, and informal problem-solving that human workers provide.
AI may complete tasks, but work is rarely just a list of tasks. It includes tradeoffs, exceptions, customer nuance, internal history, and accountability when things go wrong.
The phrase “AI psychosis” sounds dramatic, but it reflects a real concern inside the current AI cycle. The term points to a kind of executive overconfidence where AI is treated less like a tool and more like a cure-all for business complexity.
This is where the AI debate becomes less about technology and more about decision-making. A company can reasonably use AI to speed up customer service, software development, research, search, document review, or data analysis. The problem starts when leadership assumes automation will work across the organization without understanding what each function actually requires.
In practice, the easiest work to describe is not always the easiest work to replace. A support agent, for example, may not only answer tickets. They may notice product defects, calm angry customers, identify recurring issues, and escalate unusual cases. A recruiter may not only screen resumes. They may read between the lines, assess team fit, and manage delicate conversations. A developer may not only write code. They may understand legacy systems, architecture, reliability risks, and business priorities.
When companies ignore that hidden labor, AI projects can look efficient on paper while creating problems elsewhere.
The backlash is not limited to employees. TechCrunch also noted that DuckDuckGo installs are rising as some users become frustrated with Google pushing AI into search. That reaction shows a wider pattern: many people are not rejecting AI entirely, but they are rejecting products that force AI into places where they want control, links, clarity, or traditional results.
Search is a useful example because the user expectation is simple. People often want to find original sources, compare information, and decide what to trust. AI-generated summaries can be helpful, but they can also feel like an extra layer between the user and the web. If the summary is wrong, vague, or too dominant, the product feels less useful.
This is one of the biggest mistakes companies can make in the AI era. They may assume users want AI everywhere because AI is the current competitive signal. But many users want AI only where it improves the experience without removing choice.
That distinction matters. AI as an optional assistant can feel powerful. AI as a forced replacement can feel intrusive.
The AI boom is also happening alongside a difficult job market in technology. TechCrunch noted that tech layoffs in 2026 are already close to matching all of 2025. That context makes every AI announcement more sensitive.
For workers, the message can feel contradictory. Companies say AI will help employees become more productive, but some of the same companies also cite AI agents as a reason to reduce staff. That creates mistrust. Even when AI tools are genuinely useful, employees may see them as part of a replacement strategy rather than a productivity upgrade.
This tension could shape workplace adoption. If workers believe AI is being used to eliminate roles rather than improve work, they may become less open about how they use it, less willing to help integrate it, and more skeptical of leadership’s claims. The result is a culture where AI adoption is high but trust is low.
That is not a healthy foundation for long-term transformation.
One of the more useful points in the TechCrunch discussion is that both sides of the AI argument may be right at the same time. AI believers are right that the technology can improve speed, reduce repetitive work, and open up new product possibilities. AI skeptics are right that companies can overstate its reliability, underestimate hidden labor, and damage products by forcing AI into every surface.
This is why the conversation should not be reduced to whether AI is good or bad. The better question is whether companies are using it with discipline.
An AI-first strategy can make sense when it starts from a real user problem, a measurable workflow, and a clear quality standard. It becomes reckless when it starts from cost-cutting pressure, investor messaging, or the fear of looking behind competitors.
AI can help a company move faster. It can also help a company make bad decisions faster.
Companies that want to avoid becoming too AI-obsessed need a more grounded approach. They should begin by identifying where AI clearly improves a process, then test whether the improvement holds up beyond demos and internal excitement.
That means measuring outcomes, not just usage. A customer support AI should be judged by resolution quality, escalation rates, customer satisfaction, and error handling. An AI coding assistant should be judged by code quality, bug rates, review time, and maintainability. An AI search product should be judged by user trust, source visibility, accuracy, and whether people can still reach the original web.
Leaders also need to involve the workers closest to the workflow before making replacement decisions. Employees often understand the messy parts of work that do not show up in dashboards. Ignoring that knowledge can turn an AI efficiency plan into an operational mess.
The strongest companies will not be the ones that automate the fastest. They will be the ones that know what should be automated, what should remain human-led, and where the two should work together.
The current AI wave has created enormous pressure on companies to prove they are moving fast. Investors want AI strategy. Executives want lower costs. Product teams want competitive features. Workers want clarity about their future. Users want technology that helps them without taking away control.
That pressure is exactly why overreaction is dangerous.
AI is becoming a major part of business infrastructure, but it is not a substitute for understanding work, customers, or product quality. Companies that treat it as a blanket solution may save money in the short term while creating bigger problems in trust, execution, and user experience.
The better path is not anti-AI. It is anti-hype. Businesses need to use AI where it makes products better, teams stronger, and operations more reliable. When AI becomes an obsession rather than a tool, the technology stops solving problems and starts becoming one.
Share your thoughts about this article.
Be the first to post a comment!
