Why AI projects fail: Blame the culture, not the tech

2025 has been the year AI began delivering value at scale – what comes next? At the start of 2025, the call was clear: this had to be the year AI moved beyond pilots and proofs of concept into delivering measurable business and societal value. Now, as we approach the final quarter, the picture is mixed. Leading organisations have used this year to align AI with business strategy; wisely strengthening their data foundations, operationalising responsibly, and ensuring their workforce is equipped to adopt AI in their roles. They are already seeing efficiency gains, sharper decision-making, and new revenue/efficiency opportunities. The majority of organisations remain at the starting gate or in experimentation mode. They struggle with: a lack of executive commitment; poor data quality and low trust in data; an inability to define and prioritise use cases with demonstrable business value; a lack of investment in the right team, structure, and technology to support operationalised AI; and a lack of understanding about the different ways AI can be used to improve business performance.   The gap between AI leaders and laggards is widening – and 2026 will amplify that divide. Looking ahead, three shifts will define the next phase of AI adoption: 1️⃣ From isolated success to enterprise-scale transformation – organisations will move from selective use cases to embedding AI deeply into core operations 2️⃣ Trust and responsible AI as competitive differentiators – those who integrate ethics, governance, and transparency into their AI programmes will build the confidence needed to scale 3️⃣ Human + AI partnership – the future is not about replacement, but augmentation. Organisations that empower people with AI will create the most resilient and innovative models   2025 proved AI can deliver real value. 2026 will determine who can scale responsibly and earn long-term trust versus those who are left behind.

The AI Reality Check: 95% of Pilots Fail to Scale We’ve all seen the hype around AI, but here’s a sobering truth: 95% of corporate AI pilot projects never make it past the testing stage. According to a recent MIT report, the issue isn’t with the AI models themselves. The real stumbling blocks? Weak integration strategies Misaligned budgets Lack of change management Workforce unpreparedness This paints a clear picture: AI isn’t just a tech challenge, it’s an organizational one. My Take: The companies that win in the AI era won’t be those chasing flashy demos, but those investing in: ✔️ Workforce upskilling ✔️ Stronger operational frameworks ✔️ Sustainable deployment strategies AI won’t transform businesses on its own. But paired with intentional strategy and people-first execution, it can. Full article here: 95% of corporate AI pilots fail to scale, MIT report https://lnkd.in/ej8BTzZ6 👉 Question for you: Do you think this high failure rate is a temporary growing pain or a sign that the AI boom is unsustainable? Let’s talk 👇

Stop chasing AI hype. Start building AI value. AI is everywhere today, and most organizations want to use it. But many get stuck in “pilot purgatory”—running small projects that look good in presentations but never scale or deliver real value. The problem is not the technology; it’s the way companies approach it. To succeed, AI adoption must be smart and scalable. That begins with business value. Instead of chasing the newest algorithm, start with a real business problem and define clear success metrics from day one. When AI is tied to measurable impact, it delivers results that matter. Scaling also requires the right foundation. A platform approach is far more effective than isolated projects. By building reusable infrastructure and adopting MLOps, organizations can deploy, monitor, and retrain models continuously—without reinventing the wheel each time. Equally important are people. AI is not only about technology; it is about change. Companies must upskill teams and create a data-driven culture, so employees embrace AI and use it in daily decision-making. Finally, responsible AI requires governance from the start. Embedding ethics, security, and compliance ensures trust and sustainable growth. #AI #ArtificialIntelligence #AIAdoption #MLOps #Innovation

The prevailing question in boardrooms is 'Which AI tool should we adopt?'. This is, respectfully, the wrong starting point. The critical issue is not a deficit of tools, but a widespread lack of organisational readiness. Investing in AI without a foundational assessment leads to failed projects and wasted resources, a scenario confirmed by numerous industry reports. Success in AI is not about acquisition; it is about preparation. We approach this challenge with a structured framework: The Data-Process-People Audit. 1. Data Readiness: We evaluate the quality, accessibility, and governance of your data. AI is only as effective as the data it learns from. 2. Process Maturity: We analyse existing workflows to identify inefficiencies. Automating a flawed process only accelerates the generation of poor outcomes. 3. People & Culture: We assess the team's skills and the organisation's capacity for change. Technology adoption is ultimately a human endeavour. Applying this framework, as we do in our AI Opportunity Audit, provides a clear roadmap for successful AI implementation, ensuring that technology serves strategy, not the other way around. What do you consider the primary barrier to AI adoption in your industry: data, culture, or another factor?

Is Your AI Pilot Failing? It may be your Change Management 🚨 95% of enterprise AI pilots fail. That’s even higher than the infamous 70% failure rate of traditional change management. Why? Because the same flaw is being repeated. The Common Flaw: Top-Down Rigidity Too many organizations approach AI the same way they’ve approached change management: a rigid, top-down “push” model that ignores how people and workflows actually function. - New systems are imposed, not integrated. - Employee sentiment is overlooked. - Real-world impact is underestimated. 👉 Klarna learned this the hard way. After replacing human agents with AI, the company quietly rehired staff within a year. The small percentage of AI pilots that succeed don’t rely on “better algorithms.” Instead, they rely on better management: - Using AI to surface real-time employee sentiment. - Adapting strategies instantly. - Building bottom-up, human-centric change. In other words: AI succeeds when it augments human decision-making, not when it replaces it. The Hard Truth AI isn’t eating change management—it’s choking on it. The real opportunity lies in using AI to create adaptive, human-first transformations. 💡 What’s your take? Have you seen an AI transformation succeed? What made it work? 👉 Learn more about how Forge Forward supports public sector innovation: https://lnkd.in/esACzNw7 #AITransformation #ChangeManagement #DigitalTransformation #Leadership #AIStrategy #FailureToLaunch

The AI Transformation Paradox: Is Your Organization Ready for the Future? Let's be honest: AI transformation is much harder than the hype would have you believe. 🚨 Despite the promise of massive productivity gains, the reality for most organizations is that bridging the gap from existing skills to true AI literacy is a major hurdle. The issue isn't the technology itself, it's the readiness of the people and the organization's ability to adapt. This is the AI Transformation Paradox. AI capabilities have already far outpaced not only user knowledge and tool proficiency, but also the organization's ability to leverage AI at enterprise scale. MIT research confirms what we're seeing: 95% of AI initiatives are failing, not because the technology doesn't work, but because companies underestimate the fundamental shift required in behaviors and culture. This isn't just a simple technology migration; it's a complete change in how people think and work. AI introduces a new way of approaching knowledge work that demands different skills, training, and organizational foundations. The chasm between what's possible and what's currently being realized is vast, and it's getting wider every day. The question isn't if your organization will transform, but if you'll be prepared to lead it. #aiintelligence

A technology is only as good as the operating model it lands in. Josh Bersin is one of the few leaders out there talking about the need to redesign jobs and reengineer work if we are going to make the most from AI. Right now, many companies are stuck at AI assistance: Co-pilot is deployed to everyone, and most use it as a helpful assistant. People are basically doing the work they have always done, just marginally faster and outsourcing meeting notes, doc summaries, and emails to AI. The infamous 'ROI' from AI investment that leaders keep seeking? It only really comes when you reach AI augmentation. When you do the hard work of redesigning jobs and workflows to make the most of both the human and AI capabilities. When AI is hard-wired into not just what you do, but how you do it - individually, in teams, and across organisations. This is why I think OD professionals are about to have their moment. We have been reengineering operating models for YEARS. We know that sustainable, successful transformation will need a holistic, systems-thinking approach to move an organisation to a new way of working. We understand we must design jobs from the bottom up and consider tasks and accountabilities if we are to design jobs for the AI era. I know there is a lot of fear-mongering saying that there will be only '5 jobs by 2030', but we must remember: the future isn't predetermined. We have time to get this right. We still deliver human-centered AI transformation. And OD practitioners - we can be the ones to lead the charge. [picture for reach - anyone else excited about OpenAI creating their own LinkedIn? Because I cannot wait to not have to try and hack an algo].

🚨 AI is not failing. Organizations are The recent MIT headline that “95% of generative AI projects fail” has been widely misread. AI itself isn’t broken—the problem lies in how enterprises adopt and manage it. 👉 Here’s why most AI pilots stumble: Budgets flow into low-ROI areas instead of high-impact use cases. Enterprise systems lack adaptability, frustrating employees with steep learning gaps. Productivity gains leak into the shadow AI economy, bypassing official workflows. Large organizations move too slowly, losing momentum. Weak governance fails to embed AI into workflows & accountability. Hype-driven projects inflate expectations, many bound to collapse. Investors are already repricing AI-exposed firms due to poor ROI evidence. 🔗 Read the full article here: https://lnkd.in/gn9xe-aq 👉 The takeaway: AI success is less about the tech, and more about adoption strategy, governance, and speed of execution.

Why 90% of Your Company's AI Initiatives Will Fail Only 1% of companies say their AI investments are mature, even though 92% plan to spend more. Organisations are pouring money into algorithms and tools, but without redesigning workflows and culture, AI adoption stalls. Successful companies treat AI not as a tool but as a new paradigm, reshaping work so it augments human capability rather than replaces it. That's why PRGRMMD uses the BCG 10-20-70 Framework: 70% People and Processes: the core driver. This is where AI impact is realised. It means reworking entire workflows so teams focus on oversight, creativity, and higher-order decisions. 20% Technology and Data: the foundation. Legacy infrastructure creates bottlenecks. To succeed, organisations must embed intelligence into systems where decisions are made, not as an overlay. 10% Algorithms: the raw power. Advanced models are table stakes. Real advantage comes when they are tailored to people, processes, and organisational context. In Reality: 38% of Australian companies aren't adopting AI 23% are stuck at the threshold National productivity is at a 60-year low True AI transformation is an organisational and cultural endeavour first, and a technical one second. The constraint has shifted from access to models to the challenge of integration. What's the biggest AI execution challenge in your organisation?

The Unexpected Bottlenecks in Real-World AI Deployment As businesses increasingly integrate AI technologies, many discover that the promises of efficiency and innovation are often slowed by unforeseen hurdles. These bottlenecks are not just technical but are deeply rooted in organizational and market realities. One of the most critical obstacles is data quality. While AI systems are robust when trained on extensive datasets, the reality is that many organizations struggle to maintain data that is clean, updated, and relevant. This often results in AI systems that underperform or, worse, misinform crucial business decisions. According to a report by McKinsey, almost 40% of time taken by data scientists is spent cleaning and preparing data. Another significant challenge is the gap in skills and understanding. Many enterprises rush into AI adoption without adequately training their workforce in the nuanced complexities of AI tools. This leads to misaligned expectations and inefficiencies, as employees are either overwhelmed by the technology or use it suboptimally. Harvard Business Review suggests that continuous training and development are vital in empowering teams to leverage AI effectively. Finally, regulatory and ethical considerations often act as substantial bottlenecks. Regulatory compliance for AI usage imposes numerous constraints that can delay deployment, particularly in highly regulated industries like healthcare and finance. Ethical considerations, similarly, are pivotal in ensuring the responsible use of AI, yet they often lengthen development timelines as companies strive to adhere to societal expectations. Considering these challenges, how can businesses better prepare for the nuanced demands of AI integration, ensuring a smoother path to innovation? #ArtificialIntelligence #DataManagement #SkillsTraining #RegulatoryCompliance #InnovationChallenges