AI in ITSM: How to approach implementation

AI has established itself as a transformative force in the workplace. However, the rapid pace of innovation (often applied without proper consideration for specific use cases) can cause unintended consequences.

These (and other) missteps are contributing to a broader AI adoption crisis. Research firm, Gartner Inc., reports that up to 85% of enterprise AI initiatives fail.

Whether the root cause lies in false expectations (of what AI can do and how it works) or a flawed implementation approach, the issue demands examination.

AI risks

The most visible examples of AI’s shortcomings tend to occur at user-facing touchpoints—particularly with virtual agents. When these tools malfunction, their errors often become high-profile public events:

• Chevrolet’s virtual agent mistakenly offered a brand-new SUV for $1, affirming it as a legally binding deal (with no “takesies backsies”).
• DPD’s chatbot famously went rogue—composing poetry and swearing at customers.
• X’s Grok agent returned politically biased responses, clashing with widely accepted societal norms.

These incidents highlight how agentic AI, when poorly designed or badly governed, can inflict real reputational damage.

Hyper-personalized harm and workplace hesitation

Insights from Accenture’s 2025 Life Trends Report reveal a growing skepticism around AI’s role in our lives and work. As automation capabilities increase, so does the potential for hyper-personalized harm.

• 52% of people have seen fake news or articles.
• 33% have experienced deepfake scams targeting their personal information or money.
• 39% have seen fake product reviews online.

Within organizations, the problem is compounded when there is a lack of strategy, governance, communication, and education:

• 75% of companies lack a clear approach for integrating generative or agentic AI in ways that deliver positive employee outcomes.
• 60% of employees express concern that AI will increase stress and burnout.
• 37% By comparison, only 37% of business leaders share that concern—indicating an empathy gap.
• 29% of employees trust their leadership to prioritize their well-being amid and AI adoption program—confirming the empathy gap.

The numbers stress the importance of intentional, empathetic AI integration. Missteps (especially around the personification of bots) can alienate employees—leading them to feel as though their value is being compared to that of machines.

Don't rush in. Planning is essential.

Gaining a full understanding of what you want to achieve and how you should approach the implementation is critical.

Rushed AI projects have a higher chance of failure—consuming time and budget without delivering the expected value. The resulting AI-skepticism backlash can make it harder to get budget and buy-in for future projects.

To avoid the pitfalls, it is essential to:

• Spend time identifying specific service management use cases where AI can help you reduce effort, accelerate work, and improve outcomes.
• Match specific AI tools to use cases—instead of trying to adapt a more general AI tool from scratch.
• Approach the implementation with a full understanding of the inherent risks, data quality requirements, and unique aspects of an AI implementation that differ from traditional IT projects.
  

Find out more about planning and implementing AI in ITSM in this report:

Find out more

• AI-powered service management