Enterprises Are Changing Their Approach to AI as the ‘Hype Phase’ Subsides

Over the course of the last decade, corporate interest in AI investments has exploded, with Reuters reporting that global corporate investment into AI had reached over $1 trillion USD between 2021-2024.

Much of this corporate enthusiasm was caused by the success of various high-profile AI chatbots, most notably Open AI’s ChatGPT. Just two months after its launch in November 2022, the chatbot had reached 100 million users, which made it the fastest-growing consumer application in history.

This lightning rise in usage caused a major stir in the corporate world, as companies started to pour money into AI projects. A political storm, brought about principally because of fears of mass automation of jobs, ensued.

However, a report published by Solvd, an AI engineering consultancy, in April has revealed that many companies have actually started pushing back on costlier and riskier AI projects.

The hype phase: what was it and has it ended?

Solvd CEO Mike Hulbert characterized the period of frantic corporate AI investment as the “hype phase”.

“The hype peaked maybe last year, there still is some hype out there but there is [now] a greater recognition that, if this is a nine-inning game, we are probably in the second, maybe the third, inning [of AI integration]”, Hulbert told Unite AI.

In that period, large numbers of companies seemed to accept a simple formula: the more automation the better.

This somewhat cavalier attitude towards AI had large-scale political and commercial repercussions. On the commercial side, the rush to invest in AI contributed to faulty integration of automated systems on a large scale: an MIT report titled The State of AI in Business 2025 found that only 5% of generative AI pilots in the corporate world were succeeding.

On the political side, narratives warning of mass job losses abounded, with prominent U.S. Senator Bernie Sanders warning last year that AI could be responsible for the loss of 100 million U.S. jobs.

The Organization for Economic Co-operation and Development (OECD), an international forum representing 38 democracies across five continents, also warned in 2023 that 27% of jobs in its member states were at risk of automation.

Such narratives still have some traction in the media landscape, exemplified by extensive media coverage of major tech companies like Amazon and Meta citing AI as a driving factor behind large-scale layoffs.

However, the Solvd report, which investigated the attitudes of 500 CIOs and CTOs at large U.S.-based companies ($500M+ ARR) towards AI integration, indicates that continued interest in AI experimentation is now coupled with an increasing willingness to scrap so-called “hype-based” AI projects.

It shows that 49% of CTOs and CIOs surveyed expect AI pilots to become “less hype-based”. Furthermore, 72% of respondents said that it was likely that at least one AI project will be scrapped in 2026 for failing to meet KPIs, with only 14% of respondents stating that their companies maintain over half of their AI projects which show poor ROI.

One example of a high-profile “hype-based” AI project falling through is Taco Bell’s attempt to automate its drive-thru service. In 2023, the Mexican-themed fast-food chain introduced a voice AI system at 500 of its U.S. locations to reduce mistakes in orders and speed them up.

Many of Taco Bell’s competitors, such as McDonald’s, Wendy’s, Dunkin and Checkers had already implemented AI ordering.

However, the technology did not have the intended effect, as it was beset by glitches, delays and misunderstandings — the bot repeatedly asked one customer what they would like to drink with their “large Mountain Dew”.

In August last year, Taco Bell announced it would be slowing down the rollout of the technology.

The shift towards more managed integration and greater workforce protections

The Taco Bell case seems to symbolize a larger trend of companies continuing investment into AI, but more closely managing its implementation and regulation. As the rush to automate for the sake of automation has subsided, regulation of AI and more sophisticated internal oversight seem to have become greater corporate priorities.

The Solvd report showed that, while only 38% of leaders reported already having formal internal oversight for AI within their companies, 100% of respondents reported that “they have begun establishing governance frameworks” to improve issues relating to oversight.

Poor AI governance has real-world consequences: an EY survey in early March revealed that 45% of technology executives reported having suffered a confirmed or suspected leak of sensitive data due to employee use of third-party generative AI tools.

However, this previously laissez-faire attitude towards governance appears to be shifting.

The PWC 2026 Global Digital Trust Insights report, which surveyed 3,887 businesses in 72 countries, found that AI had become the top cybersecurity investment priority for respondents, as companies seek to prevent data leaks and security breaches associated with their own and third-party AI tools.

This greater AI regulation has happened hand-in-hand with greater workforce protection against potential automation-related job losses; the 2026 AI Adoption and Risk Survey – which polled 1,250 businesses from around the world — revealed that 62% of employers were actively providing their employees with on-the-job AI training to mitigate the risk of AI-related layoffs.

Hulbert described these changes positively; “I believe that there is an incredible amount of capability in the technology [AI] … [but] companies with a primary thought model of replace the human and fully automate … get into big trouble”, he concluded.