Evaluating AI: Successes, Failures, and the Investment Landscape 

As artificial intelligence (AI) technologies become increasingly integrated across industry sectors, executive leadership and private equity investment committees are under growing pressure to discern where real ROI lies. The market is navigating a critical inflection point: stories of groundbreaking AI successes are often punctuated by examples of costly missteps. With so much capital flooding the AI sector—Goldman Sachs estimates $200 billion in potential global AI investments by 2025—the fundamental question remains: Who is actually making money with AI, and where do the risks outweigh the rewards?

This article explores key considerations around AI ROI, high-profile success and failure cases, and what executive teams and investment boards need to know to navigate the evolving AI investment landscape.

AI Successes: Where AI is Delivering Tangible Results

Notable examples suggest AI can provide measurable returns when strategically implemented. IBM's recent quarterly results highlighted sustained gains in its AI and automation divisions. Their watsonx platform sees rapid adoption in sectors such as software development and customer service, with demonstrable gains in productivity and cost efficiency. IBM CEO Arvind Krishna emphasized that AI investments are now impacting hiring plans and productivity targets: “Every 1% of productivity savings on $2 trillion in tech and customer service spending is significant.” This indicates a clear path to improved margins when AI is applied accurately and systemically.

Other AI successes include Salesforce’s deployment of Einstein AI to accelerate sales pipeline development, and Amazon’s use of AI in logistics and inventory management—both contributing to measurable bottom-line growth. These outcomes underline that, when executed within a clear strategy and solid governance, AI investments can yield high ROI.

AI Failures: Lessons from Unrealistic Expectations and Costly Deployments

Conversely, AI failures highlight the risks associated with overinvestment without an operational alignment. Forrester Research reports that up to 85% of AI projects either fail to deliver on their promises or do not get deployed at all. These "proof-of-concept purgatories" often result from inadequate data infrastructure, poorly defined success metrics, or an overreliance on vendor-driven hype.

Even legacy incumbents with deep technical expertise can falter. Some internal AI projects—championed without adequate business justification—have cost tens of millions of dollars while delivering minimal end-user or client impact. In executive discussions, there's a growing recognition that not all AI experiments need to scale, and governance plays a pivotal role in recognizing when to stop or pivot technology deployment.

Who is Making Money with AI?

The real beneficiaries of the AI boom—at present—appear to be technology vendors and infrastructure providers. AI-enabled cloud platforms and GPU hardware suppliers (e.g., Nvidia) have seen exponential growth fueled by enterprise demand for computational power and AI toolkits.

Startups with focused AI applications (e.g., fraud detection, legal automation, or medical imaging) are gaining traction, particularly those with narrow, monetizable use-cases that solve real-world problems. Private equity firms see competitive advantage when investing in platforms where AI augments core services rather than replacing human workflows wholesale.

However, there’s a disparity between investor expectations and portfolio-level returns. Top-performing funds are selectively placing AI investments within operational value-building strategies rather than speculative transformations, gaining meaningful edge in portfolio company valuations.

Stephen Klein has a great LinkedIn post doing a deep dive on who makes money from AI: https://www.linkedin.com/posts/stephenbklein_please-note-i-fell-down-the-rabbit-hole-activity-7324623407336054784-TvtV/

The Evolving Metrics of AI ROI

Traditionally, AI ROI has been evaluated via cost savings or increased operational throughput. However, best-in-class organizations are now including:

  • Time-to-value: How quickly a model goes from deployment to measurable impact.
  • Risk-adjusted returns: Accounting for data quality, compliance, and reputational risks.
  • Scalability potential: Whether the applied AI system can extend across use cases or business units.

AI success metrics are becoming more nuanced, needing close alignment with business KPIs, rather than generic platform adoption metrics.

IBM has a study with more details on AI ROI: https://www.theregister.com/2025/05/06/ibm_ai_investments/?td=rt-3a

What Executive Leaders and Investment Committees Need to Know

For C-suite leaders and private equity teams, navigating the AI landscape demands inquiry beyond vendor promises. Key considerations include:

  1. Business Integration Over Technology Demos: AI investments should support broader business strategy, not exist as standalone experiments.
  2. Partnership Ecosystem: Success often depends on partnerships across system integrators, consulting firms, and vendors—not in-house development alone.
  3. Governance and Ethical Alignment: Leaders must establish robust oversight structures to ensure AI is used responsibly, particularly in regulated sectors.
  4. Human Capital Readiness: The AI workforce paradigm shift is real. Effective upskilling and role redesign are central to long-term value capture.

Final Thoughts

AI is not a monolithic investment. While some are achieving significant ROI through disciplined strategy and execution, others are navigating the sobering costs of poorly aligned initiatives. For executive leadership and private equity committees, the stakes are elevated. Understanding both the mechanics of AI success and the root causes of AI failure is essential for managing risk while uncovering real competitive advantage.

The future of AI investments depends not only on the technology’s capabilities, but on how intelligently leadership steers these investments toward value—not just novelty.

For more insights into how AI can help or hinder your company or investment strategy - contact us today.