With the push to deploy generative and agentic solutions, a common question has emerged: “Is there an AI bubble and will it burst soon?”
For many organizations, this new wave of generative and agentic AI is still in the experimental stage. The main focus is on the low hanging fruit and internal. Most companies are turning to AI to improve efficiency, such as automating workflows and streamlining customer support. The problem is that these benefits are difficult to obtain.
Ben Gilbert, vice president of 15gifts, points out that “these benefits often take years to see real benefits and are difficult to measure beyond time savings.”
This is where the cracks start to appear. The rush to unfold is uncomfortably familiar and can trigger PTSD-like feelings in some people.
“The tendency for companies to jump right into AI projects and solutions reflects a pattern we’ve seen many times in past tech bubbles, such as the dot-com era,” Gilbert explains.
This gap between experimental spending and measurable returns is precisely where bubbles are weakest.
Gilbert argues that AI projects that “focus on efficiency gains and deliver opaque or delayed ROI” will be the first to fail when the bubble bursts. Exit is inevitable when an investment “risks becoming a costly experiment rather than a profitable tool.”
“Budgets may tighten, startups may shut down, and large enterprises may reevaluate their AI strategies,” Gilbert says.
This is a data-backed warning. Gartner has already predicted that “more than 40% of agent AI projects will fail by 2027 due to rising costs, governance challenges, and lack of ROI.”
So what separates costly experiments from viable AI strategies that can survive the bubble burst? Gilbert suggests it’s because of human nuances. This is something that many projects overlook in their rush to automate. He points out that there is a strange contradiction. “Why is AI being adopted across the board for efficiency improvements and customer support, but not for sales?”
The answer may be that while algorithms are extremely valuable for sifting through data to inform decision-making, consumers also want the involvement, intuitiveness, and fluidity of human interaction. Therefore, success is about augmenting talent, not replacing it.
Gilbert argues that “AI should be taught by real humans so that it understands the nuances of human language, needs, and emotions.” This requires a transparent process, and “human annotation of AI-driven conversations allows us to set clear benchmarks and improve platform performance.”
The AI bubble is not imminently bursting completely. Gilbert explains that there is likely to be a “correction rather than a complete market collapse” and that the potential for AI remains strong. But the hype will deflate.
For corporate leaders, moving forward requires a return to first principles. “Whether AI projects are built on hype or business value, they must address real human needs to be successful,” Gilbert says.
Whether it’s a bubble or a healthy market correction, this cooling-off period may even be a good thing, as it provides an opportunity for companies to focus on the quality of their AI over hype and smarter ethics. Gilbert believes that for CIOs and CFOs managing their budgets, the brands that grow will be those that “use AI to augment human capabilities, rather than automating human capabilities.”
“Without empathy, transparency, and human insight, even the smartest AI is doomed to fail.”
See also: Keep CALM: New model design could solve high AI costs for enterprises
Want to learn more about AI and big data from industry leaders? Check out the AI & Big Data Expos in Amsterdam, California, and London. This comprehensive event is part of TechEx and co-located with other major technology events such as Cyber Security Expo. Click here for more information.
AI News is brought to you by TechForge Media. Learn about other upcoming enterprise technology events and webinars.
