What are obstacles to successfully implementing AI?

In 2023, Gartner found that 79% of business strategists considered the use of AI to be crucial for their company’s success in the next two years. However, AI had only been integrated into daily operations in 20% of cases. It turns out that successful implementation of AI hinges on two things: (1) what AI can actually do – which is improving rapidly – and (2) the willingness of people to genuinely integrate AI into their work.

In the February 2025 issue of the Harvard Business Review, Julian De Freitas, Assistant Professor at Harvard Business School, describes five obstacles to successful implementation and how to address them. What he describes is fully aligned with what we at Hi Automation experience during our implementations:

  1. AI is too opaque – a “black box.”
  2. AI has no emotions
  3. AI is not sufficiently flexible
  4. AI is too autonomous
  5. People prefer interaction with people

On top of that, there is the general concern that the introduction of AI will come at the expense of jobs. That there is a growing risk of data being used with malicious intent. And that roughly half of people believe a day will come when AI will attack humanity. Implementing AI within this general context is already a major challenge. And on top of that, the aforementioned obstacles must also be addressed in any specific company implementation.

That humans must be at the center of AI implementation and continued use is therefore evident. Being in close contact from the start and deeply understanding the underlying emotions and needs is crucial. Addressing these effectively is no easy task. An explanation in the context of an obstacle may be logically sound in itself, but may actually have the opposite effect and increase resistance.

This post is meant to point out these issues to you, not to delve into them further. The subject is too complex for that. The HBR article gives you an initial, somewhat deeper impression. Together with our subsidiary Hi automation, we at Axisto Group have the knowledge and on-the-ground experience and are happy to support you further.

Additional reading: How to achieve sustainable change at speed?

Related cases

Performance varied greatly, and that had to stop

One of the world’s largest steel producers was midway through a five-year programme to perform consistently at world-class level. At its largest plant, the hope was that Six Sigma would eliminate the remaining process variation. The team discovered that half of that variation came from well-intentioned interventions by the operators themselves. The pilot delivered far…

Growth swamped the innovation department

Growing demand for new product variants and engineering changes swamped the innovation department of a fast-growing electronics company. Design errors only came to light in production, crash actions pushed aside projects that were on schedule, and hiring more engineers made things worse. Within six months of starting work with Axisto, output was reliable again, and…

The new system was actually costing the engineers time

It turned out that the Engineers had to perform a lot of repetitive, manual, clicking, copying, and pasting that suit RPA well. The implementation of RPA has saved 5.5 man-years — and the savings still continue to grow.

Related insights

Why so many Industry 4.0 projects stall before they deliver real value

Industry 4.0 projects rarely stall because of technology. They stall because organisations approach digitalisation as a technology trajectory, while real value only emerges when strategy, operations, governance, leadership and behaviour change alongside it. That is the core of the problem. And precisely why many CEOs and COOs see that the digital ambition in their organisation…

AI and Agentic Automation in 2026

From Pilots to Industrial-Scale Performance AI is no longer the constraint. Execution with clear ROI is. Across industrial organisations, AI pilots have multiplied, tools have matured, and investment continues to grow. Yet many leaders see a familiar pattern: promising experiments that don’t translate into sustained operational performance, because they never scale into the processes, roles,…

Europe’s ESG Pay Pivot: Between Purpose and Pragmatism

Europe isn’t abandoning ESG leadership—it’s refining it. The continent’s strong regulatory foundations, from CSRD to SFDR, remain intact. But implementation is now being tempered by calls for flexibility, competitiveness, and industrial relevance. This isn’t about ESG versus growth. It’s about ESG for resilient growth. And that’s not idealism. That’s smart governance.