AI Embedded in Just 10% of Core Processes

New research reveals AI is widely used by employees, but rarely integrated into the workflows needed to deliver real business value

Australian organisations are rapidly adopting artificial intelligence (AI), but most are not applying it within the core processes that run their business, according to new research from Appian [Nasdaq: APPN]. While more than half of surveyed employees report using AI within their organisation, only 10% say it is embedded into core business processes, where work is executed, decisions are made, and outcomes are measured.

Nearly one in three employees also say their organisation still relies on standalone AI tools rather than integrating AI into end-to-end workflows, undermining its ability to deliver scalable, meaningful results.

Luke Thomas, Vice President, Asia Pacific and Japan at Appian, said the findings highlight a fundamental gap in how organisations are approaching AI adoption.

“AI is being introduced across organisations, but often not in the processes that actually run the business. Just 10% of Australian workers say AI is embedded into core business processes, which lags behind global findings from a recent Harvard Business Review Analytic Services study, sponsored by Appian, in which 18% of respondents reported that AI is primarily integrated within workflows,” Thomas said. “Process is how organisations structure work, make decisions and serve customers. When AI isn’t connected to that, it lacks the context, control and visibility needed to deliver meaningful outcomes, and importantly, to measure its impact.”

AI is visible, but not deeply embedded

The research shows Australian employees most commonly report AI being applied in customer service (20%) and marketing (17%), with lower adoption in core operational areas such finance (8%) and HR (6%).

This suggests organisations are prioritising more accessible, lower-risk use cases, rather than embedding AI into the operational workflows that drive business performance.

“Adoption alone isn’t the issue when it comes to delivering value from AI, it’s application. Many organisations have introduced AI, but until it is embedded into core processes, where work actually happens, it becomes difficult to drive consistent improvements,” said Thomas.

Expectations remain high, but outcomes are mixed

While AI adoption is accelerating, many organisations remain in the early stages of translating that momentum into value, with expectations often outpacing what employees are seeing in practice.

More than 46% of employees describe their organisation’s expectations of AI as optimistic or overly optimistic, yet only 23% report significant improvements. A further 11% say they have seen no measurable impact, while 15% say it is too early to tell.

“Organisations are understandably optimistic about what AI can deliver, but the results are still varied. In many cases, employees are seeing isolated improvements rather than consistent gains across day-to-day work, which makes it harder to build long-term confidence,” observed Thomas.

Integration and skills gaps remain key barriers

Employees identify a range of technical and organisational barriers to scaling AI.

Skills and talent gaps were cited as the biggest challenge (27%), followed by integration with existing systems (20%). Governance concerns (14%) and lack of clear strategy (13%) also remain significant factors.

These findings reinforce the need for a more structured approach to AI adoption, where AI is fully integrated with existing systems, data, processes and guardrails.

Focus shifts to full AI integration

Encouragingly, organisations are beginning to prioritise deeper AI integration.

The most common focus over the next 12–24 months is integrating AI into core systems and processes (32%), ahead of expanding existing use cases (19%) or continuing experimentation (17%).

“AI has enormous potential, but its value becomes clear when it’s integrated into the processes that run the business and applied to real operational challenges. This is the case for the National Injury Insurance Scheme Queensland (NIISQ), which supports individuals seriously injured in motor vehicle accidents on Queensland roads. NIISQ uses Appian’s gen AI capabilities to extract and identify fields for invoice processing with over 80% data extraction accuracy. AI has delivered strong results for NIISQ with a 12x ROI and reduced manual effort by 50%,” said Thomas.

“Australian organisations risk limiting AI’s impact if they adopt it in a piecemeal way. Adding tools to solve isolated tactical problems won’t translate into meaningful performance gains. That requires applying AI across core business activities.”