Why 95% of AI Projects Are Failing (and What’s Actually Working)

17/10/2025 7 min
Why 95% of AI Projects Are Failing (and What’s Actually Working)

Listen "Why 95% of AI Projects Are Failing (and What’s Actually Working)"

Episode Synopsis

The AI revolution isn't playing out as expected. While tech headlines herald the transformative potential of artificial intelligence for business, a surprising reality is emerging: 95% of enterprise-level AI implementations have failed to deliver financial returns within six months, according to recent MIT research.Meanwhile, small to medium enterprises are quietly succeeding where giants stumble. Between 30-40% of SMEs report achieving meaningful cost efficiencies from their AI investments. What explains this counterintuitive outcome? The answer lies in approach and execution. Large organizations typically pursue custom-built solutions requiring significant investment and organizational change, while smaller businesses leverage ready-made tools that solve specific problems immediately.The most successful implementations aren't replacing creative functions but optimizing repetitive tasks—automating finance operations, streamlining data entry, and reducing time spent on menial work. This pattern suggests today's AI is best suited for process optimization rather than completely replacing specialized human roles. We've seen this reality check play out at major companies like Commonwealth Bank, where an AI call center implementation initially backfired, increasing call volumes and necessitating the rehiring of human staff.Looking ahead, AI implementation is becoming increasingly accessible through platforms like N8N and Make.com, which create low-code environments where even non-technical users can build powerful automations. This democratization will likely accelerate adoption among smaller businesses that can quickly implement and iterate without massive investments. However, as businesses grow more dependent on these AI platforms, we should expect subscription costs to rise significantly—following the pattern of other essential business tools.What repetitive processes in your business could benefit from AI automation today? The opportunity is clear: start small, focus on specific pain points, and leverage existing solutions rather than reinventing the wheel.The Digital Horizons Podcast is hosted by:James Walker - Managing Director WHDBrian Hastings - Managing Director Nous