If you feel like you have whiplash when it comes to AI, you’re in good company. First, McKinsey & Company called 2023 the “breakout year” for generative AI (GenAI). The technology has dominated tech conversations — and news headlines — ever since. Then, with companies in practically every industry embracing the technology and deploying it in their operations, Newsweek in July 2024 predicted that the AI bubble was about to burst.

So, which is it? Is AI the technology that’s going to change the world, or a shiny object that didn’t live up to expectations?

The answer is probably somewhere in between. Whatever happens, though, embedded in the last two years of AI hype is an important lesson for companies that are considering new and emerging technologies: It’s not necessarily that companies should be more cautious or conservative about investing in new solutions; rather, it’s that they should be more strategic.

An AI bust similar to the dot-com crash of the early 2000s is unlikely. Whereas dot-com companies lacked viable business plans and offered no real value, AI’s rapid rise is fueled by fundamental technological advancements that have real promise across industries, according to Matthew Gilbert, a marketing lecturer at Coastal Carolina University.

“During the dot-com bubble, the emperor had no clothes; with AI, he now has an entire wardrobe,” Gilbert said. “While there will likely be consolidation among companies offering AI services, the need for them will only continue to grow.”

Is Gartner’s Hype Cycle Hyping a Bust?

Many point to Gartner’s hype cycle as proof that AI is over hyped. But the final stage of the famed model — following the “Peak of Inflated Expectations” and the “Trough of Disillusionment” — is often overlooked: The so-called “Plateau of Productivity” is full of sustained and practical benefits.

For companies that are worried about an AI bust, the question isn’t how to avoid it, but rather how to prepare for it by hastening the journey to the promised land that is the “Plateau of Productivity.” Because value from AI will largely be found in projects with standardized processes using familiar AI techniques, Gartner in its most recent Hype Cycle update recommended that AI leaders look at composite AI techniques.

“GenAI still has potential to be a transformational technology with profound business impacts on content discovery, creation, authenticity and regulation, automation of human work, and customer and employee experiences,” Gartner noted. “Still, GenAI faces challenges, including ethical and societal concerns, limited security best practices and nefarious uses like deep fakes and disinformation.”

Challenges of Implementing AI

Although Gartner remains bullish on AI, its research concedes that 30% of GenAI projects will be abandoned after proof of concept by 2025. Reasons include poor data quality, inadequate risk controls, escalating costs and unclear business value. In most cases, however, the biggest reason for abandoning GenAI projects is the investment required and the challenge of quantifying the ROI, Gartner reported.

Clearly, AI projects are rife with challenges. Those who focus on the 30% of GenAI projects that will be abandoned, however, are overlooking the important reality that 70% of GenAI projects — the vast majority — will continue on the path toward implementation. While many of those projects will bear fruit, companies must be patient waiting for their harvest.

“GenAI requires a higher tolerance for indirect, future financial investment criteria versus immediate return on investment,” Gartner reported. “Historically, many CFOs have not been comfortable with investing today for indirect value in the future. This reluctance can skew investment allocation to tactical versus strategic outcomes.”

Along with patience, success requires realism, according to Debanjan Saha, CIO of DataRobot. While GenAI models will continue to improve, he predicts that they eventually will hit a ceiling with regard to capabilities. After all, most of the written text that’s available has already been used to train models, which leaves audio and video as the next frontier.

“When you lift a cup from a table … nobody writes it down,” Saha told The Forecast. “But there are a lot of videos which show how we go and grab the cup and pick it up from the table. Those are the kind of data sources which are now going to make these models more and more intelligent. [But] I do think there is a limit to which you can feed them information and make them better.”

Selecting the Right Use Cases

If concerns about an AI bust are rooted in the challenges of implementing and operationalizing AI, organizations that want to win at AI must focus on practical, high-ROI use cases.

A common error, for instance, is using AI for use cases that require human intervention. Consider Coca-Cola’s “Create Real Magic” campaign, where the company invited customers and artists to use AI tools like GPT-4 and DALL-E to create digital art for the campaign. In short order, Coca-Cola received criticism for its AI-made holiday ad, a remake of the classic 1995 “Holidays Are Coming” ad.

“Critics called the ad ‘unnatural,’ ‘soulless’ and ‘lacking emotional depth,’ claiming it did not show the sincerity and warmth of the original,” noted Gilbert, who said the ad illustrates “the difficulties of using AI in creative areas where human input is essential.”

Customer support automation and deflection is the clear leader for companies wanting quick wins, according to Dev Nag, CEO of AI startup QueryPal. He said it’s the perfect use case because companies can quickly implement and also clearly measure results. Many businesses see both cost savings and improved customer satisfaction within just a few weeks.

“Modern AI can handle a significant portion of routine inquiries while actually providing better service than traditional automated systems, and it gives existing agents more time to focus on strategic accounts,” Nag said. 

Predictive maintenance is another promising use case that can quickly generate value. Instead of waiting for machines and equipment to malfunction, organizations use AI to monitor their performance in real time and predict problems before they occur. Businesses can then plan for maintenance or switch to another machine, preventing disruptive and costly downtime.

“Look for processes that involve lots of routine document processing, analysis or decision making,” Nag said. “These are areas where AI can augment human capabilities effectively without requiring massive organizational changes, and they often have the best existing data for training. The key is starting with well-defined, bounded problems where success can be clearly measured, current spend is well understood and historical data is plentiful.”

Getting the Most Value from AI

So, is AI a bubble waiting to burst, or a juggernaut that will continue to get bigger and stronger over time? 

“I think we're all getting AI fatigue,” Steve McDowell, chief analyst at NAND Research said in a video interview with The Forecast. 

“With a lot of these technology transitions, whether it's the internet, whether it's the smartphone, whether it's the PC, going way back, there's a brief burst of time where we're very excited about the technology pieces, but the value of technology comes in how I use it. So have we hit a wall on AI? No, but I think it's time to change the conversation. Let's stop talking about GPUs. We'll stop talking about who the providers are, and let's start talking about what are we going to do with it and how's it going to change my business.”

He said as organizations continue making investments in AI, they will increasingly focus on aspects that create real business value. To do that, organizations will need to get clear about how they will measure success. 

McDowell said IT practitioners will lean on software to manage across different infrastructures. Innovation in this area is causing a lot of disruption and it’s changing “the way that we think about infrastructure,” said McDowell.

“I like GPT-in-a-Box that the Nutanix is delivering because that's a software set of capabilities,” he said. “I can deploy that at the edge or I can roll that out in the cloud, if that makes sense.

Enterprises also need to develop internal capabilities, whether that's training existing staff or hiring new talent, according to Nag.

“You need people who understand both the technology and your business domain,” Nag said. “The organizations that are succeeding with AI right now are the ones taking a thoughtful, measured approach focused on specific business outcomes rather than chasing the latest buzzwords or trying to transform everything overnight.”

Jennifer Goforth Gregory has written for Microsoft, Adobe, IBM, Google, Salesforce, Verizon and AT&T. In her spare time, she rescues more than 130 homeless dachshunds each year and finds them new forever homes.

Ken Kaplan and Jason Lopez contributed to this story.

© 2025 Nutanix, Inc. All rights reserved. For additional information and important legal disclaimers, please go here.