IT leaders are being stung by "unexpected" AI costs

IT leaders are struggling with unexpected AI-related costs, along with legal, reputational, and compliance issues.

In a poll conducted by Asana, more than eight-in-ten UK-based IT leaders said they’d encountered unplanning AI cost increases over the last 12 months.

Just over six-in-ten said they are held highly or fully accountable for AI-driven business outcomes . However, with AI adoption increasingly happening across departments and outside traditional governance processes, those responsible for AI ROI are unable to fully track and measure its usage.

The survey by Asana comes amid rising concerns over the skyrocketing costs associated with AI, particularly agentic AI.

Latest Videos From

Watch full video here:

As ITPro reported earlier this month, Uber was hit with an eye-watering bill following a push to drive AI adoption and use internally at the ride-hailing giant. Indeed, the company is believed to have blown through its entire annual AI budget in just four months.

This incident, as well as the rise of ‘tokenmaxxing’ across the incident, has prompted some enterprises (including Uber) to introduce caps on AI token use and rethink how productivity with the technology is measured.

AI costs aren't the only worry

Elsewhere in the Asana survey, governance was identified as a particular headache among IT leaders.

More than half (53%) said that an AI tool or agent has taken an action in the past 12 months that resulted in financial, legal, reputational, or compliance harm.

Nearly half of UK workers say they're not aware of how their company uses their work – including emails, documents and tasks – to train or improve AI systems. One-in-four say that they often use AI tools that are not formally approved by their organization, while 38% regularly use personal AI accounts for work tasks.

UK IT leaders estimate, indeed, that three-in-ten new AI tools and agents that have been introduced over the past year have been adopted without formal IT review or approval.

"Most organizations are past asking whether their people will use AI – they already are. The challenge now is turning that into measurable business value, without losing the governance and visibility needed to manage risk," said Christina Francis, head of UKI and Northern Europe at Asana.

"People want faster, smarter ways to work. The question is whether organizations can give them trusted, approved ways to do that, or whether they end up working around the tools entirely."

Stalled AI projects still plague IT teams

The research also revealed problems with execution, with 46% of UK IT leaders saying that AI initiatives often or always fail or stall because AI lacks complete organizational context.

Meanwhile, 37% of UK knowledge workers spend at least 30 minutes a day fixing or reworking AI outputs due to missing context – a trend also highlighted in a recent study by Glean.

Workers are also struggling to access the information AI needs to perform effectively: four-in-ten say they need to check three or more tools to gather the context needed to complete their work, while two-thirds say they have to re-explain context to AI tools at least some of the time.

"AI is most powerful when it has context: the goals, decisions and workflows that sit around the work. But that only lands if there's transparency built in. Employees need to understand how their work feeds AI systems and what that makes possible," said Francis.

"The organizations that get ahead will be the ones that combine visibility, governance and context not as competing priorities, but as the same thing."

FOLLOW US ON SOCIAL MEDIA

Follow ITPro on Google News and add us as a preferred source to keep tabs on all our latest news, analysis, views, and reviews.

You can also follow ITPro on LinkedIn, X, Facebook, and BlueSky.