The World Has Updated Its Expectations. Have You Updated Yours?

A Global Inflection Point in AI Adoption

A seismic shift has occurred in global artificial intelligence adoption. In 2023, people were exploring the technology out of curiosity; by 2024, they began to see its practical benefits; and by 2025, AI had become an integral part of daily life and workflows. Google and Ipsos’s third annual “Our Life with AI” report—based on interviews with 21,000 people across 21 countries—quantifies this transformation: two-thirds of the global population now report using an AI tool. Personal virtual-assistant usage has climbed 24 percentage points since 2023, reaching 62%.

Yet the real change lies not in the numbers but in how people use the technology. Last year, the vast majority engaged with AI for entertainment or exploration. In 2025, the picture is different: users turn to AI to grasp complex topics, save time, get input before making difficult decisions—even to navigate major life changes. The technology has evolved from a curiosity into a tool.

This shift raises a critical question for organizations. As customers and employees develop increasingly mature relationships with AI, are enterprise AI systems meeting those expectations? At CBOT, we hear this question with growing frequency across projects in diverse industries—and the data suggest that the right time to build a sound enterprise AI strategy is not “soon” but now.

In this article, we examine the report’s key findings through an enterprise AI lens.

Adoption Is Now Universal

Last year, roughly half the global population said they used AI—a significant milestone at the time. By 2025, that share has reached 66%; personal use of virtual assistants rose 24 points from 2023 to 62%. The report characterizes this as crossing an “adoption threshold.”

For enterprises, the implication is clear: the customer who interacts with AI is no longer the exception—they are the norm. Whether you look at banking, retail, or telecommunications, the person on the other side of the conversation has at least some AI experience. We see this vividly in CBOT projects: users now approach virtual assistants with less patience, higher demands, and less forgiveness—because they can compare.

The demographic profile is also shifting. In 2024, AI usage skewed heavily toward younger, more educated, higher-income individuals. By 2025, the gender gap has narrowed from 10 points to just 3; parents and teachers have emerged as power users. This broadening reach signals that enterprises must design AI solutions not for a narrow segment but for a far wider and more diverse audience.

Usage Has Matured

The 2024 report showed that most people used AI primarily for entertainment and exploration. In 2025, that pattern has reversed. Seventy-four percent of users engage AI to understand complex topics, 70% to support their work, and 65% to save time. Even more striking: 40% have consulted AI while making a major life decision, and 42% have used it to explore a new career or business idea.

AI is no longer a toy; it is a thinking partner. This transformation represents both an opportunity and a warning for enterprises.

The opportunity: Customers now attach greater meaning to AI interactions—a well-designed virtual assistant can build a genuine value relationship. The warning: Shallow systems that merely answer FAQs no longer meet expectations.

We see this on the ground at CBOT: organizations have moved rapidly from “we have a bot” to “how effective is our bot?” The benchmark has changed. As usage deepens, expectations rise. A year ago, users were discovering what AI could do; today, they know what it should do. That distinction is decisive for enterprise AI design.

Excitement Is Growing—but So Is Workforce Anxiety

In the 2024 report, excitement about AI overtook apprehension for the first time—57% excited versus 43% concerned globally. In 2025, the balance still tilts toward excitement, yet the picture is more nuanced. The share who view AI’s transformation of jobs as “a good thing” has dropped 9 points year over year. Public opinion on workforce impact is now evenly split: 50% believe AI will help workers, while 50% expect job losses.

It is important, however, to recognize that anxiety is not uniform. Among those who actively use AI at work, optimism is markedly higher—63% in this group say AI will be helpful. Younger generations exhibit a similar pattern. Concern is concentrated among those with little or no first-hand experience. This trend, visible in earlier reports, has sharpened in 2025: experience reduces anxiety.

For enterprises, this finding has a practical implication. The success of AI projects is measured not only by technical performance but also by how much employees trust these systems and how comfortable they are working alongside them. At CBOT, we observe this in project after project: AI deployments that fail to center the employee experience eventually generate resistance. Adoption must be won continuously, not once.

Trust Is Stable, but Expectations Are Rising

Trust in AI companies has barely moved over three years: 55%, 56%, 57%. Although this plateau may appear static, it tells an important story—the public has not yet been disappointed, but neither has it embraced these companies enthusiastically. Trust has not been earned; it simply has not yet been lost. That is a subtle but critical distinction.

Trust in governments, by contrast, has eroded slightly since 2024. In the West, citizens are uncertain whether their governments are using AI actively and effectively—four in ten Europeans still believe regulation is overly restrictive. In emerging markets, the picture is the opposite: both trust and expectations are high, and there is demand for governments to do more. The fact that two worlds relate so differently to the same technology shows that AI requires context-sensitive strategies rather than one-size-fits-all solutions.

The same dynamic operates on the enterprise side. The more time users spend with AI systems, the narrower their tolerance becomes. “Not bad” is no longer good enough—systems that behave inconsistently, miss context, or promise more than they deliver erode trust. There is one window to build trust; far fewer to rebuild it once lost.

Experimentation Is Over—Now Comes Strategy

When three years of data are read in sequence, a single direction emerges: AI is entering the lives of more people, more deeply, every year. The journey that began in 2023 with “Does this actually work?” has evolved by 2025 into “How do I use this better?” The experimental phase has closed—we have entered the strategy phase.

This transition creates both pressure and opportunity for enterprises. The pressure: Users are now experienced with AI. They can compare, articulate their expectations clearly, and recognize disappointment quickly. The tolerance window for superficial solutions is closing. The opportunity: The ground has never been more prepared for systems that are well-designed, genuinely integrated into operations, and capable of generating measurable value.

Perhaps the report’s most significant finding is this: those who use AI more are less anxious, more optimistic, and hold higher expectations. This feedback loop applies to the enterprise side as well—systems that deliver the right experience build trust, trust drives deeper engagement, and deeper engagement produces tangible value. The starting point is always the quality of that first experience.