Event Summary | Why 80%+ of AI Projects Fail – and What Organisations Need to Do Differently

The Swedish Australian Chamber of Commerce recently hosted an insightful online event exploring one of the most pressing challenges facing organisations today: Why do so many AI projects fail despite rapid technological advancement and growing investment?

The session featured AI and analytics expert Evan Shellshear, who shared insights from extensive research conducted together with co-author Doug Gray for their recent book examining the success and failure of AI initiatives across industries.

Drawing on more than 1,000 peer-reviewed articles, industry interviews, and real-world case studies, the presentation highlighted a sobering reality: over 80% of AI, analytics, and data science projects fail — and organisations are not significantly improving over time.

A central theme throughout the discussion was that AI failure is rarely caused by the technology itself. Instead, the biggest barriers are often organisational and human factors, including poor strategic alignment, lack of leadership buy-in, weak data quality, unrealistic expectations, and resistance to organisational change.

Key Takeaways from the Session

• AI success is primarily a people and strategy challenge, not a technology challenge
  The session highlighted that culture, leadership, governance, and organisational readiness are far more important than the AI tools themselves.
• Analytical maturity is a key differentiator
  Organisations with clear AI strategies, leadership support, strong data culture, and the right processes significantly outperform less mature organisations.
• AI projects must align with business strategy
  Successful organisations focus on AI initiatives that solve real business problems and directly support strategic priorities — rather than launching AI projects simply because of market hype.
• Data quality remains one of the largest barriers to success
  Poor-quality or inaccessible data continues to be one of the most common reasons AI initiatives fail across industries.
• AI transformation requires change management
  Implementing AI successfully often requires organisations to rethink workflows, operating models, and decision-making processes — not simply introduce new tools.
• Small wins build momentum and trust
  Organisations were encouraged to begin with achievable AI projects that demonstrate measurable value and support longer-term transformation.
• Leadership buy-in is essential
  Executive and board-level support is critical to ensure AI projects receive the necessary funding, strategic alignment, and organisational backing.
• Emerging opportunities in Agentic AI
  The discussion also explored the growing impact of “Agentic AI” in areas such as finance, administration, HR, and back-office operations, where organisations are already seeing significant productivity gains.

The session concluded with a strong reminder that while AI technologies continue to evolve rapidly, long-term success will depend on organisations’ ability to combine technology with strong leadership, culture, governance, and people capability.

Thank you to Evan Shellshear for sharing his expertise and research, to moderator Aldo Gutierrez for leading the discussion, and to everyone who joined the session and contributed to the conversation.

About the speakers:

Dr Evan Shellshear is currently a Principal at BCG X (Boston Consulting Group’s AI and Data Science division). He has a BA and BSc from the University of Queensland, a Masters and PhD from the Nobel Prize winning Institute of Mathematical Economics at the University of Bielefeld in Germany and over fifteen years of experience in building global AI and statistical algorithms for industries from manufacturing to retail, HR to supply chain, oil and gas to energy. He has published or co-published four books on topics ranging from the Amazon Best Seller Innovation Tools to AI in Medicine as well as the Wiley best seller Why Data Science Projects Fail: The Harsh Realities of Analytics and AI, without the Hype. He is currently an Adjunct Professor at the University of Queensland and Queensland University of Technology, having taught topics from business analytics to AI strategy.

Link to book https://www.amazon.com.au/Why-Data-Science-Projects-Fail/dp/1032660309