AI Adoption Risks in Organisations: Why the Failure Isn’t the Technology

By Alexandra Egan, MICDA Strategic Review & Improvement Consultant

AI is not the problem. Uncontrolled adoption is.

That distinction matters more than most organisations currently recognise. Across sectors, from small and medium enterprises to not-for-profits to large corporates, the same pattern keeps appearing. Tools are being used. Results are inconsistent. No one is quite sure who is responsible. And the risks are quietly accumulating.

A recent ICDA and McKinsey guide found that 76% of not-for-profits have no gen AI strategy, and 80% lack an official use policy. Those numbers are striking. But they are not surprising. And they are not unique to the NFP sector.

The pattern is familiar across every type of organisation. The technology arrives before the structure does.

The Four Risks Nobody Is Talking About

When AI adoption goes wrong, it rarely fails because the tool did not work. It fails because the organisation was not ready for it. There are four risk patterns that appear consistently, regardless of industry or size.

1. Teams are confused

When there is no clear direction on how AI should be used, individuals make their own decisions. Some use it constantly. Others avoid it entirely. The result is inconsistency across outputs, processes, and client experience. Confusion is not a technology problem. It is a leadership problem.

2. Standards are unclear

What does a good AI output look like in your organisation? What level of review is required before it is used? What quality standard must it meet? Without answers to these questions, the quality of work becomes dependent on the individual rather than the organisation. That is a risk that compounds over time.

3. Ownership is missing

When something goes wrong with an AI output, and eventually, it will, who is accountable? If the answer is unclear, the organisation has a governance gap. Accountability cannot be assigned after the fact. It must be designed in from the beginning.

4. Risk is not always visible

This is perhaps the most dangerous of the four. Executives and boards are not always positioned to see the risks that are building quietly beneath the surface. Shadow AI use, inconsistent outputs, ungoverned data handling, these do not announce themselves. They surface as incidents, complaints, or compliance failures.

This Is a Leadership and Operating Model Issue

Organisations often frame AI adoption as a technology challenge. It is not. The technology is, in most cases, accessible, functional, and improving rapidly.

The real challenge is organisational. It sits in the operating model, the governance framework, the culture, and the leadership approach. Until an organisation deliberately addresses these four areas, it will continue to experience the same pattern of uneven, ungoverned AI use.

Think of it this way. If your organisation can’t answer the following four questions clearly and consistently, chances are, you do not yet have an AI strategy. You have variation.

Where should AI be used in our organisation?
Where should it not be used?
What standard must outputs meet before they are acted upon?
Who is accountable when something goes wrong?

What Australian SMEs and Organisations Are Telling Us

The data on Australian AI adoption reflects this gap clearly. According to the National AI Centre Adoption Tracker and the MYOB Business Monitor (2025):

Approximately 37% of Australian SMEs are currently adopting AI; meaning the majority are not yet using it in any structured way
42% of Australian SMEs have no plans to adopt AI at all
46% of SMEs already using AI do not measure its impact. no return on investment tracking, no accountability framework
23% of SMEs remain unaware of AI’s potential applications for their business

These are not technology gaps. They are strategy, structure, and governance gaps. And they are solvable.

The Diagnostic: Is Your Organisation Ready?

Before scaling AI adoption, every organisation should be able to answer five foundational questions. These form the basis of a structured AI readiness review.

1. Do you have an AI technology policy? Not a vague statement of intent. A clear, practical policy that defines what AI can and cannot be used for, what tools are approved, and what the process is for introducing new ones.

2. Do you have a leadership and operating model for AI? Who owns AI in your organisation? Who sets the direction? Who resolves disputes about use? Without clear leadership, AI governance defaults to whoever happens to be most enthusiastic and that is not a strategy.

3. Do you have an organisation design that supports AI? Your structure should reflect your AI ambitions. If AI is expected to improve efficiency, reduce duplication, or support decision making, the organisation needs to be designed around that, not retrofitted after the fact.

4. Do you have a culture that supports responsible AI use? Culture determines whether policies are followed or ignored. Organisations with strong cultures of accountability, quality, and continuous improvement tend to adopt AI more successfully. Those without them tend to amplify their existing problems.

5. Are you optimising with intention? Where is automation genuinely appropriate? Where does human judgement remain essential? Answering these questions deliberately rather than letting the technology decide, is what separates strategic adoption from uncontrolled variation.

Variation Eventually Becomes Risk

The cost of uncontrolled AI adoption is not always immediate. It accumulates. Inconsistent outputs erode quality. Ungoverned use creates compliance exposure. Missing accountability frameworks leave organisations vulnerable when things go wrong.

And when risk is left unmanaged, it becomes exposure to reputational damage, to regulatory scrutiny, to the erosion of trust with the clients, funders, and stakeholders your organisation depends on.

The failure is not in the absence of effort. Organisations are trying. Leaders are curious. Teams are experimenting.

The failure is in the absence of structure. And structure is something that can be built.

Where to Start

If your organisation is in the early stages of AI adoption or if you suspect your current approach has gaps, the most valuable first step is an honest assessment of where you actually are.

Not where you would like to be, but where you are, right now!

That means looking at your policies, your governance, your operating model, and your culture with clarity and without defensiveness. It means asking the questions above and sitting with the answers, even when they are uncomfortable.

The organisations that will use AI well are not necessarily the ones that move fastest. They are the ones that build the right foundations first.

Alexandra Egan is a Strategic Review and Improvement Consultant specialising in strategy, structure, transformation, and operational improvement. She works with SMEs, not-for-profits, and complex organisations to build the foundations for sustainable performance.

To find out more or to discuss a strategic review for your organisation, visit alexandraegan.com.au

Sources: National AI Centre Adoption Tracker, Department of Industry, Science and Resources (2025); MYOB Business Monitor (2025); ICDA/McKinsey Not-for-Profit AI Guide (2024)