AI fails when people are left behind

Most discussions about AI implementation begin in the wrong place.

They focus on models, platforms, and technical capabilities. Organisations debate vendors, evaluate tools, and measure performance against predefined metrics. Yet despite growing investment, failure rates remain stubbornly high. Some estimates suggest that between 85 and 95 per cent of AI projects fail to deliver meaningful value.

The explanation may be simpler than many organisations would like to admit.

Successful AI adoption is rarely a technology problem. More often, it is a human problem.

This is the theme of Episode 10 of The AI Values Podcast, where Edosa Odaro and Lindley Gooden speak with Amy Shi-Nash, Professor of AI Practice at Monash University and Co-Founder & CEO of Hippo Labs. Drawing on more than 25 years of experience implementing AI across large organisations, Amy argues that the difference between successful and unsuccessful deployments often comes down to adoption, culture and organisational readiness rather than technical sophistication.

Success begins when AI becomes part of daily work

Many organisations define AI success through technical milestones. A model is deployed. A pilot is completed. A system moves into production. While these achievements matter, Amy suggests they are only the starting point.

The more meaningful question is whether people actually use the system.

"Using these AI systems has become a daily routine or daily habit and it's part of their job." 

A successful implementation does not exist. It becomes embedded in everyday work. Employees use it consistently, teams incorporate it into decision-making, and the technology becomes part of normal operational behaviour rather than a separate innovation initiative.

This distinction is important because adoption reveals something deeper than usage statistics. It demonstrates whether people find the technology genuinely useful.

For boards and executive teams, this changes how success should be measured. Technical performance remains important, but it is not enough. Leaders also need visibility into behavioural adoption, workforce engagement, and whether AI is becoming part of the organisation's operating model.

The real value emerges after adoption

One of the most interesting observations in the discussion is that value often appears later than organisations expect.

Many AI programmes begin with a specific use case. The objective is usually to improve efficiency, automate a process, or provide quicker access to information. These benefits matter, but they are not necessarily where the greatest value is created.

According to Amy, the more significant shift occurs when people begin rethinking how work itself should be done.

"Once you get a feel of what AI could do, you actually think differently." 

As employees become familiar with AI, they start identifying new opportunities, redesigning workflows and creating solutions that were not anticipated at the beginning of the project. Teams move beyond using AI as a tool and begin collaborating with it as part of the problem-solving process.

"That's where the value really starts to evolve and expand because you're not forcing AI into the human design." 

This is where many organisations underestimate the potential of AI. They focus on efficiency gains while overlooking the possibility that AI may fundamentally change how value is created.

The implication for leaders is clear. Successful implementation is not simply about deploying technology. It is about creating the conditions for experimentation, learning and continuous improvement.

Culture change is the real transformation

Throughout the conversation, Amy repeatedly returns to a theme that is often overlooked in AI strategy discussions: culture.

When people learn how to work alongside AI, they develop new habits, new skills, and new ways of approaching problems. Over time, these behaviours become embedded within the organisation.

"That new skill is basically allowing the workforce to think how to co-work with an AI system or co-create something. And once that becomes a norm, that actually means organisational culture change."

What begins as technology adoption gradually becomes cultural change.

This matters because culture determines whether innovation scales. An organisation may deploy an impressive AI system, but if employees remain reluctant to experiment, challenge assumptions or explore new possibilities, the technology will struggle to deliver its full value.

Amy describes successful organisations as those where continuous learning becomes normal. Employees become comfortable adapting their workflows, developing new capabilities, and exploring how AI can support their work.

"Your imagination has expanded and your skills continue to learn."

The result is not simply improved productivity. It is an organisation that becomes more capable of adapting to future change.

For leaders, this represents a significant governance challenge. Culture cannot be installed like software. It must be developed intentionally through leadership, incentives, and sustained engagement.

Readiness is about more than technology

One of the strongest sections of the discussion focuses on organisational readiness.

Many organisations spend considerable time debating whether AI is ready for them. Amy suggests that leaders should perhaps be asking a different question: are they ready for AI?

"As an organisation, are we ready? And how do you know we're ready or not?"

Readiness extends beyond infrastructure and technical capability. It includes organisational mindset, workforce literacy, competitive pressures, regulatory expectations and the willingness to experiment. These factors often determine whether an organisation can translate AI capability into measurable value.

The conversation also highlights an important tension. While discussion and debate are essential, there is a risk in remaining permanently in analysis mode. Organisations that spend too long observing developments may miss opportunities to learn through practical experience.

Amy argues that conversation is essential, but conversation alone is not enough. Organisations must create opportunities to experiment, validate ideas and learn through doing rather than remaining indefinitely in observation mode.

This creates a delicate balance for leaders. They need enough governance to manage risk, but enough experimentation to generate learning. Organisations that achieve this balance are often better positioned than those pursuing either extreme.

Why resistance is not the problem

AI adoption is often framed as a battle between supporters and sceptics. The reality is more nuanced.

"Not everyone is gung ho... Some people are nervous, uncertain."

Amy notes that people respond to disruption differently. Some focus on opportunities, while others focus on risks. Some want immediate action, while others prefer a longer time horizon. Differences in literacy, confidence and experience also shape how individuals engage with emerging technologies.

Rather than treating these differences as obstacles, organisations should recognise them as valuable inputs into decision-making.

"What can we help to bring people together is the important part."

Healthy debate helps organisations identify blind spots, anticipate risks and develop more balanced approaches to implementation. Problems arise when discussions become polarised into extremes, where organisations feel pressured to be entirely committed to AI or entirely resistant to it.

The lesson is not that everyone must agree. The lesson is that organisations need mechanisms for bringing different perspectives together.

Successful adoption often depends less on eliminating tension and more on managing it productively.

AI is changing the shape of work

The conversation concludes by exploring one of the most significant long-term consequences of AI: the transformation of work itself.

Amy argues that AI will not simply create or remove jobs. It will change the composition of jobs, the structure of organisations and the skills that people need to succeed. Many tasks currently performed by humans will increasingly be supported by AI, while human expertise becomes concentrated around judgement, creativity, problem-solving and domain knowledge.

This shift has implications far beyond workforce planning.

Organisations may become flatter as access to information becomes easier and decision-making becomes more distributed. Specialised expertise may become more valuable as routine work becomes increasingly automated. New forms of entrepreneurial and fractional work may emerge as barriers to entry continue to fall.

At the same time, lifelong learning becomes essential.

"One of the skills that's definitely become more important in the era of AI is that lifelong learning habit."

The traditional model of education followed by decades of stable employment is already under pressure. In its place emerges a world where individuals continuously develop new skills and adapt to changing technological realities.

"It doesn't matter where you are... you can be the graduate, you can be a CEO or you can be a board member. Everybody is facing this change."

For leaders, the challenge is not simply preparing for technological change. It is preparing people for a different relationship with work itself.

The future belongs to organisations that learn

Episode 10 ultimately presents a different perspective on AI success.

The conversation is not really about models, platforms, or technical architectures. It is about whether organisations can create the conditions for people to adapt, learn and participate in change.

Technology may enable transformation, but people determine whether transformation succeeds.

The organisations most likely to realise value from AI will not necessarily be those with access to the most advanced systems. They will be those who build cultures of learning, create space for experimentation, and ensure that people remain at the centre of implementation.

If AI is reshaping how work is done, then leadership's role is not simply to deploy technology.

It is to ensure that nobody is left behind.

Listen to Episode 10

Going deeper: why AI readiness is a leadership capability

Many organisations evaluate AI readiness through a technical lens. They assess data quality, infrastructure, security, and governance frameworks. These factors remain important, but Episode 10 suggests they represent only part of the picture.

Readiness is increasingly becoming a leadership capability.

Organisations must determine whether employees understand the technology, whether leaders can communicate a clear vision, and whether teams have the confidence to experiment responsibly. Without these foundations, even technically successful implementations can struggle to generate meaningful adoption.

This importantly reframes AI governance. Responsible AI is not solely about managing technological risks. It is also about creating organisational conditions where people can engage with change constructively.

The organisations that thrive will be those that treat learning, adaptation and workforce participation as strategic assets rather than afterthoughts.