After three years of experimentation and spending, companies are starting to demand results as talk of an AI bubble looms. According to Kyndryl’s recent Readiness Report, based on insights from 3,700 business executives, 61% of CEOs say they are under more pressure to demonstrate a return on their AI investments compared to a year ago.
As AI development continues to move forward at breakneck speed, business leaders are being challenged to balance long-term innovation with the need to prove results now. It also creates the risk of misalignment among the C-suite, with technology and business leaders looking after their companies’ innovations while finance leaders look after their balance sheets.
“Last year we had a lot of experimental budgets. It was like, ‘Let’s give every department a budget and experiment with whatever tools they find useful,'” said Lexi Reese, a former strategy executive at companies such as Google and current CEO of AI observability platform Lanai. “This is responsible acceleration, because this price is very expensive.”
Spending on AI skyrockets
It has been widely reported throughout the year that unprecedented amounts of money are being spent on the development and deployment of AI. Much of this has to do with infrastructure spending and eye-popping startup investments by Frontier AI Institute, but corporations are also investing heavily. Garner predicts that spending on AI application software will more than triple from last year to nearly $270 billion in 2026. Reese said he discussed the cost of AI tools with more than 300 customers over the past year and found they were spending between $590 and $1,400 per employee per year, according to internal data shared with Fortune.
For many executives, this is bringing back flashbacks to past digital transformations, especially moving to the cloud, and for some, it leaves a sour taste in their mouths. Michael Bradshaw, global practice leader for applications and data at Kyndryl, told Fortune that he consistently recognizes this in his role of working with executives to bridge the gap between business strategy and what they’re actually trying to do with technology.
“Nearly two-thirds of respondents said their cloud strategy came about by chance, and 95% of them said that if they could do it all over again, they would do it again,” he said, citing data from a Kindrill report. “This is shocking to me as a practitioner because the same thing is about to happen to us, and I think we’re seeing it today with AI.”
For example, he pointed to one customer who spent more than $1 billion just implementing their first ERP (enterprise resource planning) software. “When they’re looking at the next wave, their frame of reference is, ‘Oh my God, on top of the difficult business environment, we’re going to have to spend the same amount of money to do the next wave of transformation. We can’t afford that!'”
Manisha Khanna, senior product manager for AI and generative AI at analytics firm SAS, told Fortune that this is the biggest challenge she hears from customers right now, asking, “How much should I invest in these technologies to get an ROI?”
“Our customers have seen very flashy demos from various vendors. They have high expectations. But then they start asking the question, ‘Do you really understand how long this will take to go into production?’ There are compute costs, data infrastructure requirements and human resources costs,” she said. “How do you budget for all that?”
Challenges hindering ROI
When Dan Rogers took over as CEO of Asana just a few months ago, AI ROI immediately became a top priority. Understanding the returns of AI is a top priority, he told Fortune, using a formal approach that includes both financial ROI and “human-centric” ROI metrics such as reduced administrative burden and improved decision-making. He said department leaders own the AI achievements in their fields and must report on specific metrics, adding that the company is “setting ambitious efficiency goals as part of a top-down strategy.”
At the same time, this raises the challenge of how to weigh long-term bets against short-term value.
“If you look for financial ROI too soon, you’ll kill the experiment. If you wait too long, you’ll end up in pilot purgatory,” he said.
Rogers said he is wary of forcing any AI effort to clear narrow short-term hurdles, as the benefits may be incremental. For example, some functions are infrastructure layers, and you need to think long-term about what that layer enables. On the other hand, AI’s rapid innovation cycle can make waiting and watching especially difficult.
“AI’s pace of change has effectively shattered traditional planning cycles. We have moved from a 12-month ABR (annual business review) cadence to quarterly checkpoints and continuous prioritization, because we can’t wait a year to course-correct in an area that evolves every month,” said Rogers.
Kyndryl’s Bradshaw agreed, explaining that the company is moving into a “very rapid cycle where we’re talking to customers about delivering value within four to six months.” Another factor Bradshaw points out is that the ROI of AI can be difficult to measure, especially now that so much of that ROI revolves around personal productivity technologies.
“You have agents helping you process your mail. How do you translate that into personal productivity and what business outcomes are you delivering? That’s why companies really struggle to articulate ROI,” he said.
All of this is adding up and cracks are starting to appear in the C-suite. According to a report from Kyndryl, nearly three in four CEOs say short-term ROI pressures hurt long-term innovation, and 65% say they disagree with their CFO on long-term value.
“CFOs might want to look at it from a balance sheet perspective. Business leaders want to make sure the business model changes. Technology leaders like CTOs want to make sure I’m innovating, that I’m applying all the latest technology, that I have the talent and skills in my team to truly deliver value,” Khanna said. “And these three (executives) are in very different positions right now in terms of expectations.”
New year of AI ROI
Asana’s Rogers believes the industry’s big conversation around spending has definitely increased the focus on ROI. “And to be honest, it’s too early,” he said. “When you hear reports that 95% of AI pilots make zero profit, boards and executives naturally ask tougher questions,” he said, adding that pursuing a clear ROI is a positive step, and that it is time to move the discussion “from novelty to results.”
The growing desire to prove ROI could lead to a boom in companies like Reese’s Lanai. Reese’s Lanai promises to help customers gain a clearer picture of how their teams are actually using AI, what’s having an impact, and what’s not. Lanai specifically analyzes all the prompts that users are entering into the AI tool to get a detailed view of their work. Reese said the key to increasing ROI is identifying the “highest value workflows” and adjusting from there.
“Once you see the work and the workflow, you can start standardizing the work and say, ‘Sales sees a lot of value in these workflows, so let’s make them even stronger. These workflows are being adopted a lot, but they’re very dangerous. Let’s point people in the right direction,'” she said.
While there will certainly be pressure to prove ROI in 2026, Khanna and Bradshaw believe most companies are not in a position to get there.
“We’re going to see that in people we call ‘pacesetters.’ So there’s alignment within the C-suite, there’s alignment with the employee base, there’s alignment on the technology strategy,” Bradshaw said. “There will be a few[companies that will achieve ROI in 2026]but it won’t be widespread. That’s my prediction, because most companies won’t be in a position to achieve that.”
