Every month, we continue to be alerted to the high failure rates of AI initiatives. Some of the highlights providing insights on some of the AI trends are summarized below:
-
Around 80–85% of AI projects fail, especially failing to reach production—a rate twice that of traditional IT initiatives. This estimate is consistent across multiple sources, including RAND and Gartner. Source: AI Curator, Informatica, and RAND Corporation
-
Gartner reports that only about 30% of AI projects make it past the pilot stage. Source: Informatica and RheoData
-
S&P Global Market Intelligence highlights a sharp increase in abandonment: In 2025, 42% of enterprises scrapped most of their AI initiatives—up from 17% in 2024. On average, organizations abandoned 46% of AI proof-of-concepts before they ever reached production. Source: CFO Dive and CIO Dive
-
Generative AI (GenAI) also faces its challenges 30% are projected to be abandoned post-proof-of-concept by the end of 2025 due to unclear value, rising costs, and data issues. Source: TechRepublic
So, why is this happening?
With my twenty-five-plus years of experience navigating mid to large enterprise digital transformations, it is not too difficult to understand why AI projects are consistently failing everywhere.
What is challenging is changing the methods of HOW organizations are implementing AI.
Simply stated, many organizations are treating AI adoption as a technical rollout rather than as a deep organizational business transformation.
I have outlined eight strategic reasons why AI is failing, highlighted gaps, and impacts
1. Lack of Governance and Operating Accountabilities
-
The Gap: A centralized governance operating model bringing diverse stakeholders together to be responsible for a clearly defined AI strategy and Strategic Roadmap (2-3 years outlook. Upgrading to AI is hard work and is not to be taken lightly.
-
Impact: Employees see that key leaders are given AI leadership roles across the organization, and policies/standards are clearly defined. A focused governance and set of operating accountabilities give employees confidence that AI is taken seriously.
2. Lack of Vision & Alignment
-
The Gap: Clear articulation of why AI is being implemented, how it supports business goals, how it fits into existing workflows, clarity of impact to shifts in job roles and responsibilities, ensuring effective communication at all development stages, and actively involving employees impacted by the new AI initiatives.
-
Impact: Employees see AI not as a threat, or as a “management fad,” rather AI is seen as a strategic enabler improving both the effectiveness and efficiencies of employee roles. Employee Engagement in all AI solutioning stages further increases employee trust in AI and builds resilience to change.
3. Minimal Stakeholder Engagement
-
The Gap: Early involvement of all impacted stakeholders (leaders, frontline teams, IT, compliance, unions, etc.) is identified at the start of the AI design process. Key stakeholders are positioned as change leaders, carefully identified, and engaged to ensure their needs are listened to. A “win-win” strategic vision is established to mitigate adoption risks.
-
Impact: AI tools are rejected or underused because people feel it was “done to them” rather than “built with them.” Humans support what they create. With over 50% of humans not trusting AI or the intentions of management’s goals in applying AI, no board or executive team can invest in AI without being more prepared to counter the high risk of not generating a positive ROI. The odds are against AI projects succeeding; hence, the more human support systems that are integrated throughout the change journey, the more likely AI implementation outcomes will be successful.
4. Insufficient Communication
-
The Gap: Transparent, ongoing messaging about benefits, limitations, risks, and progress. Clear communication on job roles and responsibilities that may shift due to AI enablements is critical. Involving stakeholders throughout all design stages is critical and having ongoing communication through all development stages of the AI solutioning is imperative. If jobs are at risk, organizations need to be transparent and plan effectively for the transition support requirements.
-
Impact: Not providing sufficient communication will accelerate rumors and fears, which will take over the narrative, especially around job displacement or surveillance. Communication must come in many forms of employee engagement: employee team communication, town hall meetings, corporate newsletters, visiting the design labs, identifying change agents to bring back the “right communication” to the organization – all these practices, and many more, should be tailored to be relevant to the transformation program’s needs.
5. Inadequate Skills & Training
-
The Gap: Robust, hands-on upskilling programs to build digital /AI literacy increase employee confidence and trust. AI learning and development is like learning a new language and building new skills; it cannot be done by an email, a memo, or a chatbot alone; organizations must invest in AI training relevant to the “jobs to be done” in the organization where AI is being applied. Providing diverse education and learning programs is a critical foundation to advance AI adoption.
-
Impact: If employees are not engaged in uplifting their core AI knowledge and digital literacy, they will actively resist. And in many cases, employees will easily revert to old processes or misuse AI outputs because they don’t understand how to interpret or challenge them.
6. Ignoring Culture & Trust Factors
-
The Gap: Alignment with organizational values, ethics, and fairness principles.
-
Impact: If AI is perceived as biased, opaque, or “un-human,” adoption rates plummet. Humans need clarity and confidence to trust AI, but when cultural values don’t align with AI initiatives, humans simply don’t support the changes. They can act like stubborn mules and dig in, resisting using the new toolkits, often because the tools don’t reflect “the jobs that need to be done”, and a lack of integrated design thinking has marginalized AI solutioning.
7. Poor Change Readiness Assessment
-
The Gap: Not investing in measuring organizational readiness (tech maturity, data quality, leadership support before deployment) before starting an AI project, so the entry pathways are thoughtfully grounded to ensure operational readiness.
-
Impact: AI launches fail due to fragile infrastructure or a culture not ready to experiment with AI, and the wrong AI use cases are identified without carefully aligned AI strategic thinking.
8. No Iterative Human in the Loop Feedback
-
The Gap: During all stages of an AI project, human-in-the-loop interactions are needed through all stages of the AI development process. Change management enablements are required to support post-launch listening, monitoring, and adjustments. AI models don’t stand still, so the change management support systems require ongoing investments and iterative human-in-the-loop feedback.
-
Impact: Early adoption stumbles are never corrected, leading to long-term disengagement.
Conclusion
AI initiatives fail not because of the algorithm but because organizations skip the “people and process” side of change management — the same mistake made with ERP, CRM, and other major digital transformations.
With the poor ROI outcomes on AI initiatives, leaders need to engage experienced AI and change transformation advisors to augment their own AI solutioning teams. The odds of failure are so high – board directors and C-level executives have to stop and THINK much harder. Delegating AI programs to the CIO or CTO, or AI Chief Data Scientist to lead without a strong governance operating model, investing in AI-relevant technology infrastructure, and change management enablements is a guaranteed sinkhole strategy.
In our customer AI projects, we often have to guide organizations to improve their data lineage and accuracy practices before AI can be applied. We have been able to improve many organizations’ CRM programs where companies have used Salesforce unsuccessfully due to poor workflow designs and misaligned employee performance metrics. (You can see some of our select customer AU case studies here)
We have been helping global organizations advance their AI journeys with strong workflow foundations and integrated change journey methods. We have developed robust AI change readiness assessments to guide our customers and identify risks so that AI value is realized, and most importantly, sustained.
