95% of AI projects fail but Sci-Fi solves AI…….AGAIN

A report, The GenAI Divide: State of AI in Business 2025, published by MIT’s NADA initiative, found that 95% of pilots stall at early stages and never progress to scaled adoption. Only 5% of projects achieved rapid revenue growth.

Based on 150 interviews with business leaders, a survey of 350 employees and an analysis of 300 public AI deployments, the study highlights a growing divide between successful AI integrations and those that remain stuck in experimental mode.

This and other surveys have found two key, and unwelcome, trends.

• The Perpetual Pilot trap, reduces risk but not cost as expensive investment often throws good money after bad, in the hope of finding something that produces a returnhe Perpetual Pilot trap, reduces risk but not cost as expensive investment often throws good money after bad, in the hope of finding something that produces a return
• Opportunities in isolation, local, cheaper, low risk AI implementations. But which fail to relate to or deliver any Strategic level benefits.

What these identify is a failure to understand how AI can relate to, enhance and even transform your business.

We should be learning from repeated history…..

Let me take you back to the 1980s. The IBM PC debuted in 1981. I first encountered it in 1986 in BT. The PC had three transformative impacts (well those that come to mind just now):

• It did away with job roles such as seeing the demise of typing pools
• new job roles were created and mushroomed such as software development
• and many roles were transformed, such as managers who changed how they worked

Those however, were the outcomes, as the PC led to transformation of business operating models.

• Strategy changed: E.g. product, IT, procurement, resourcing,   
• Organisation structure: e.g. new and expanded departments and relationships e.g. IT departments and cross-functional projects
• People changed: new and changed roles and behaviours, office environments
• Processes: mostly changed but some new and some gone, office facilities etc.
• Technology: The IBM PC then networking and and and…….

Some thoughts on AI integration……..or perhaps they are principles

• AI is a business solution NOT a technology solution
• Don’t think AI led, think People led
• Don’t think AI driven…..think AI supported
• To benefit from AI, think holistically and,
• Make sure you REALLY understand how your organisation works – think of it as an iceberg - before you try and change it.

• Make sure you understand Change Management, and its delivery vehicles, Portfolios of Programmes and Projects. C level folk who do not ‘get’ projects need to wise up VERY quickly, if you want to be in the successful 5% and not the failure 95%.

And a final thought. While you are holistically transforming how your organisation works, consider that AI/Digitisation and Agility have goals and outcomes in common.

• Goals: Cost reduction and revenue growth
• Outcomes: Efficiency, adaptability, flexibility, behaviours…….

Agile organisations are more profitable.

P.S. A little reminder about my planned 1 day workshop on what most project management course do not teach……Stranger in a Strange Land - how to make your projects thrive in an unfriendly organisational landscape. Planned for 30th April 2026 in London just after the EVA33 conference. Full details shortly

https://www.garudax.id/pulse/30th-april-workshop-stranger-strange-land-how-help-projects-pyne-wotbe

P.P.S. My book https://practicalinspiration.com/book/agile-beyond-it