Most ANZ firms still fail to scale AI beyond pilots

Adobe has published research showing that most organisations in Australia and New Zealand have not scaled artificial intelligence across their businesses. The study found that only 5.2% qualified as AI Leaders.

The report was based on a survey of 465 senior executives in Australia and New Zealand with authority over AI investment, alongside interviews about adoption challenges and opportunities.

A central finding was that organisational structure, rather than access to models or spending alone, was the strongest indicator of whether AI moved beyond pilots and early deployments. Many businesses had launched initiatives across teams, but progress often slowed when they tried to expand those projects across the wider organisation.

According to the research, 59% of organisations are in what Adobe describes as the "Frozen Middle", meaning they are piloting or deploying AI but lack the coordination needed to scale it. A further 45% are still working on their first use case or stopped after an unsuccessful attempt.

The figures point to a small group making broader progress. Just 23 organisations in the dataset met Adobe's definition of AI Leaders, which required both scaled deployment across multiple workflows and an operating model designed to sustain it. Those organisations converted pilots into production at nearly four times the rate of others.

Timeframes also appear to be slowing adoption. Seven in ten organisations said it takes at least six months to move from experimentation to initial deployment, while almost a third said the process takes 12 months or more.

Where projects stall

Only about 15% of proposed AI ideas make it into production. Around 34% fail at the IT evaluation stage because business teams often lack the data and architectural grounding needed to define implementation properly. Another 18% are rejected during security review, with that rate rising sharply in heavily regulated sectors including financial services, healthcare and life sciences.

The issue is not a shortage of ideas, but a lack of repeatable internal processes to move them into live use. These can include approved integration patterns, standardised security reviews and pre-defined budget categories, reducing the need for each project to go through a new process.

The survey also pointed to persistent labour market and internal training problems. Eighty per cent of respondents said AI skills shortages had either stayed the same or worsened over the past 12 months. At the same time, about half of organisations are broadening existing roles to include AI work, while a quarter are adding new skill requirements.

Operating model gap

The report found that many organisations scored more strongly on strategy and governance than on structure and workflows. Adobe argued that approving a policy or strategy is easier than changing reporting lines, ownership, decision rights and day-to-day working methods.

Among the leaders identified in the survey, maturity in structural settings was 61% above the regional average, while workflow maturity was 48% higher. These organisations were more likely to have clear ownership, accountability and deeper integration of AI into routine business processes.

Ownership emerged as one of the clearest dividing lines. Only 16% of organisations in Australia and New Zealand had a designated AI owner with budget authority and accountability for outcomes, compared with 96% among AI Leaders.

Measurement was another weak point. Just 10% of companies said they measured AI return on investment comprehensively, with most tracking usage rather than output quality, rework or the cost of checking AI-generated material.

The report also found that 79% of surveyed companies still treated AI delivery like a ticketing process, in which business units submit requests and IT responds. By contrast, leading organisations tended to involve business and IT together before a project brief was written.

Jeremy Willmott, Senior Director, Digital Strategy Group APAC, Adobe, said: "We expected the barriers to broader AI adoption to centre on technology, and depend on which models are being used, how much spend is allocated, or how early companies begin their AI adoption journey. Instead, the strongest predictor of whether AI scaled is organisational readiness."

He added: "After all, two companies can run the same models, spend the same money, and hire similar teams, yet our research suggests that one can compound its return on investment while the other stalls. The difference comes down to the operating model, including who owns AI, how it's measured, how business and IT work together, and how the work itself gets redesigned."

Despite the difficulties, spending is still rising. The study found that 78% of companies are increasing AI investment year on year, while almost two-thirds believe they are still underinvesting.

Willmott said: "What we found was that companies often lack pre-approved pathways that help move initiatives into production, including approved integration patterns, standardised security reviews, and established budget categories. Without these mechanisms, proposals must navigate a new or different process, significantly reducing their chances of success."

On workforce constraints, he said: "This remains a widespread and stubborn challenge, with 80% of businesses saying the skills shortfall remained unchanged or worsened over the past 12 months."