42% of Companies Abandoned AI Projects in 2025 — The 5-Factor Checklist That Prevents It

The Abandonment Crisis No One Predicted

Something alarming happened in 2025. The rate at which companies abandoned their AI projects did not just increase — it surged. According to S&P Global's Voice of the Enterprise survey, 42% of companies abandoned the majority of their AI initiatives before reaching production, up from just 17% the prior year (Pertama Partners, AI Implementation Success Factors for Asian SMBs, 2026). That is a 147% increase in a single year.

To put it bluntly: nearly half of all organizations that committed budget, talent, and organizational energy to AI walked away with nothing to show for it.

This is not a maturation cycle. It is not the normal growing pains of an emerging technology. It is an acceleration of failure during a period of record investment — global enterprise AI spending exceeded USD 300 billion in 2025. For small and medium businesses in Southeast Asia and Hong Kong, where a single failed AI project can consume an entire year's technology budget, the implications are existential.

The question every SMB leader should be asking is not whether to pursue AI. The competitive imperative is clear. The question is: how do you avoid becoming part of the 80% that fail?

The answer, based on meta-analysis of 2,500+ global AI implementation projects and direct pattern analysis from 50+ SMB engagements across Southeast Asia, lies in a structured pre-implementation diagnostic. Not better technology. Better preparation.

Why AI Projects Fail: The Real Root Causes

The instinct when an AI project fails is to blame the technology. The algorithm was not accurate enough. The platform was not mature enough. The vendor oversold and underdelivered.

But the data tells a different story. BCG research found that technology accounts for only 10% of whether an AI initiative succeeds or fails, with the remaining 90% determined by data foundations, people, and processes (Pertama Partners, AI Implementation Success Factors for Asian SMBs, 2026). Pertama Partners' own analysis confirms that 84% of AI implementation failures in Asian SMBs are attributable to leadership, organizational, and process factors rather than technical limitations.

The failure taxonomy, drawn from RAND Corporation research and corroborated across multiple studies, breaks down into five recurring patterns:

• Misaligned problem definition — organizations start with a technology looking for a problem rather than a problem seeking a solution
• Data infrastructure gaps — 63% of organizations either do not have or are unsure whether they have the right data management practices for AI
• Organizational resistance and change failure — despite 88% of organizations claiming regular AI usage, only 5% of employees say they use AI to genuinely transform their work
• Inadequate leadership engagement — when AI is delegated entirely to technical teams, projects lose strategic alignment
• Scope miscalibration — attempting enterprise-scale transformations before proving value at unit economics

Every one of these causes is diagnosable before a single dollar is spent on technology. That is the purpose of the Pertama 5-Factor AI Success Model.

The Pertama 5-Factor AI Success Model: A Pre-Implementation Diagnostic

The model identifies five interdependent conditions that must be present for an AI implementation to succeed. It is not a sequential process — it is a system. Weakness in any single factor undermines the others, while strength across all five creates a compounding effect.

Organizations that systematically address all five factors achieve a 3.2x higher success rate than those pursuing technology-first approaches (Pertama Partners, AI Implementation Success Factors for Asian SMBs, 2026).

Here is how each factor works as a diagnostic checkpoint.

Factor 1: Leadership Alignment

The single most powerful predictor of AI implementation success is not the quality of the technology — it is the engagement of the CEO. McKinsey research demonstrates that when the CEO or board takes direct oversight of AI initiatives, the result is a 3.6x boost in bottom-line impact compared to initiatives delegated to technical teams alone (Pertama Partners, AI Implementation Success Factors for Asian SMBs, 2026).

In Asian business cultures, this effect is amplified. Hierarchical endorsement is not merely helpful — it is structurally necessary. CEO-endorsed AI projects see 18% higher adoption rates than bottom-up initiatives, but only when paired with structured change management.

Your checklist questions:

• Can the CEO articulate the specific business problem AI will address in one sentence?
• Is budget approved, ring-fenced, and protected for at least 12 months?
• Has an executive sponsor been named with AI delivery in their performance objectives?
• Has the leadership team completed at least 4 hours of AI literacy training?

If you cannot check all four, you are not ready to select technology. Investing in leadership alignment first costs weeks; skipping it costs months and hundreds of thousands of dollars.

Factor 2: Data Readiness

Gartner predicts that through 2026, organizations will abandon 60% of AI projects unsupported by AI-ready data (Pertama Partners, AI Implementation Success Factors for Asian SMBs, 2026). This is not a future risk — it is a current reality. Only 12% of organizations report data of sufficient quality and accessibility for effective AI implementation.

For Asian SMBs, data readiness is simultaneously the most underestimated and most addressable factor. Many organizations assume their existing data — customer records, financial reports, operational logs — is sufficient. It rarely is. But SMBs, with smaller and more centralized data environments, can achieve data readiness faster than enterprises burdened by decades of legacy systems.

Your checklist questions:

• Is core business data digitized and centralized in structured systems?
• Has a data quality audit been completed within the last 6 months?
• Are data governance policies documented and enforced?
• Does sufficient historical data exist for the target use case?

The most cost-effective AI investment an unprepared organization can make is not an AI platform. It is a data cleanup and centralization project that creates the preconditions for any future AI initiative.

Factor 3: Change Management

Here is the statistic that should reshape how every SMB leader thinks about AI: despite over USD 300 billion in projected enterprise AI spending and nearly 90% of knowledge workers claiming to use AI at work, only 5% of employees say they are using AI to genuinely transform their work (Pertama Partners, AI Implementation Success Factors for Asian SMBs, 2026). The gap between organizational claims and individual reality is not a reporting anomaly. It is an implementation chasm.

Change management is consistently the most underinvested factor in Asian SMB AI implementations. SMBs that invest at least 15% of their total AI project budget in change management achieve 2.4x higher user adoption rates at 90 days post-deployment.

Your checklist questions:

• Is there a written communication plan for all affected employees, including explicit job security messaging?
• Has a training program been designed with protected time allocation (minimum 2 hours per week for the first 90 days)?
• Have incentive structures been reviewed and modified to reward AI adoption behaviors?
• Are internal AI champions identified (2-3 per affected team) and trained?

Deploying AI tools to an unprepared workforce is the organizational equivalent of installing a commercial kitchen in a restaurant with no trained chefs.

Factor 4: Government Funding Navigation

This factor is unique to the Southeast Asian and Hong Kong context and represents a structural advantage unavailable in most Western markets. Government funding programs can reduce the direct financial risk of AI implementation by 40-70%, fundamentally changing the ROI equation.

The data makes the case: government-funded AI implementations in Southeast Asia show a 67% completion rate versus 31% for self-funded projects of equivalent scope (Pertama Partners, AI Implementation Success Factors for Asian SMBs, 2026). The difference is attributable not only to reduced financial pressure but also to the structured planning that funding applications require.

Programs like Singapore's SkillsFuture Enterprise Credit, Malaysia's HRDF/HRD Corp digital training grants, Indonesia's GenAI Open Innovation programs, and Hong Kong's Digital Transformation Support Pilot Programme exist specifically to de-risk SMB AI adoption.

Your checklist questions:

• Have relevant funding programs been identified for your jurisdiction?
• Are eligibility requirements confirmed and documented?
• Is the application timeline aligned with the project timeline?
• Does the organization have internal capacity to manage funding compliance and reporting?

Organizations that navigate funding successfully have, by virtue of the application process, already addressed critical planning gaps.

Factor 5: Right-Sizing

Enterprise AI approaches fail at SMB scale. The most dangerous AI strategy for an SMB is copying what large enterprises do.

Pattern analysis reveals a stark divide: 78% of successful AI implementations in Asian SMBs started with a use case costing under USD 50,000, while 71% of failures began with projects exceeding USD 200,000 (Pertama Partners, AI Implementation Success Factors for Asian SMBs, 2026). The relationship between initial scope and success is not linear — it is binary. Small starts succeed; large starts fail.

Your checklist questions:

• Is the first use case specific, measurable, and achievable within 90 days?
• Does the technology solution match internal technical capacity (buy vs. build)?
• Is the total first-project budget under USD 75,000?
• Is a single AI Project Owner designated with at least 20% time allocation?

Asian SMBs that complete implementation within 90 days of project initiation are 3.1x more likely to achieve their target ROI than those with longer timelines. Speed does not mean rushing — it means right-sizing scope so that execution stays focused.

How to Use This Checklist: The Scoring Approach

Each of the five factors can be scored on a 20-point scale, producing a total readiness score out of 100. The scoring is not academic — it directly predicts outcomes:

The checklist is not a gate designed to prevent AI adoption. It is a diagnostic designed to ensure that when you do invest, the investment produces returns rather than joining the growing ledger of abandoned initiatives.

The Cost of Skipping the Checklist

The average Asian SMB loses USD 247,000 per failed AI implementation when accounting for direct costs, opportunity costs, and organizational trust erosion (Pertama Partners, AI Implementation Success Factors for Asian SMBs, 2026). But the financial loss is not even the most damaging consequence.

Organizations experiencing a high-profile AI failure take 18-24 months to regain internal appetite for AI investment. During those 18-24 months, competitors that implemented successfully are compounding their advantage — the first successful AI deployment shortens the timeline for the second by 60%. The gap between organizations that start right and those that start wrong does not close over time. It widens.

The Compound Effect of Getting It Right

The flipside of the failure data is equally compelling. SMBs that address all five factors and implement successfully create organizational momentum that accelerates everything that follows:

• 91% of SMBs using AI report revenue growth, with positive ROI achieved within 6 weeks on average
• The first project builds internal capability, establishes governance, and creates a cohort of AI-literate employees
• The time from first successful deployment to second deployment is typically 60% shorter
• Within 12-18 months, a right-sized, structured approach can build meaningful AI capability across the organization

The question is not whether SMBs can succeed with AI — the data shows they can, and often faster than enterprises. The question is whether they prepare properly before they begin.

Read the Full Research

For the complete framework including all five factors, case studies, and the full readiness checklist, read AI Implementation Success Factors for Asian SMBs.

The research paper includes detailed scoring rubrics for each factor, ten anonymized implementation case patterns from across Southeast Asia and Hong Kong, a complete decision framework with 100-point assessment, and the full ROI calculation methodology.

Take the First Step

The checklist approach works because it forces the right conversations before money is spent. It surfaces gaps that are far cheaper to address in planning than in production. And it gives leadership teams the confidence that their AI investment is built on a foundation of evidence, not optimism.

Ready to implement AI the right way? Book a consultation with Pertama Partners. We will walk you through the full 5-Factor readiness assessment tailored to your organization, your market, and your specific business challenges.