UK IT leaders face rising AI costs and governance gaps

An Asana poll of UK IT leaders and knowledge workers found that more than 80% of IT leaders experienced unplanned increases in AI costs over the past 12 months. Fifty-three percent reported an AI tool or agent had caused financial, legal, reputational or regulatory harm in the previous year. Respondents estimated roughly 30% of new AI tools were adopted without formal IT review.

Just over 60% of IT leaders said they feel highly or fully accountable for AI-driven business outcomes, even as AI use spreads outside central IT teams. The survey flagged gaps in visibility when tools are introduced without formal approval, creating blind spots for those managing risk and return on investment.

One in four employees reported they often use AI tools that are not formally approved by their employer, and 38% said they regularly use personal AI accounts for work tasks. Nearly half of workers said they do not know how their emails, documents or tasks might be used to train or improve AI systems.

Respondents linked rising costs to higher model use and to so-called agentic AI. Heavy consumption of model tokens, described in the poll as “tokenmaxxing,” contributed to higher bills. Some firms reported imposing caps on token use and reviewing how they measure productivity gains from AI.

More than half of respondents reported governance failures that led to harm in the last 12 months. The poll did not provide aggregated monetary totals for those harms.

IT leaders also reported execution problems. Forty-six percent said AI projects often or always fail or stall because AI lacks complete organisational context. Thirty-seven percent of knowledge workers said they spend at least 30 minutes a day reworking or correcting AI-generated results for missing context.

Employees said they frequently must consult multiple systems to give AI the context it needs: 40% check three or more separate tools to gather information, and two-thirds said they must re-explain context to AI tools at least some of the time.

Christina Francis, head of UKI and Northern Europe at Asana, stated that most organisations already have staff using AI and urged clearer controls and visibility. She added that AI performs best when it has access to the goals, decisions and workflows around tasks and called for greater transparency so employees understand how their work feeds AI systems.