A minority of businesses have won big with AI. What are they doing right?

  a-minority-of-businesses-have-won-big-with-ai-what-are-they-doing-right.html
  Business and Finance on Tue, Oct 14th, 2025 by ZDNet

Only a Minority of Companies Are Reaping the Full Rewards of AI—and Here’s Why

Despite the flood of hype surrounding artificial intelligence, recent data reveal a stark reality: only a small fraction of businesses have translated AI promises into tangible gains. A 2023 study by Gartner found that a mere 8 % of organizations achieved the ROI levels they had anticipated, while the majority struggled to move beyond pilot projects. The ZDNet article “A minority of businesses have won big with AI: what are they doing right?” dissects the factors that set the successful few apart and offers a playbook for the rest.

1. Strategic Alignment Matters

The article opens with a compelling observation: companies that integrate AI into a clear, business‑driven strategy fare far better than those treating it as a technological buzzword. Successful firms start by answering two critical questions:

1. Which business outcomes can AI realistically influence?
2. How will AI fit into the organization’s broader mission and KPIs?

The piece cites the example of Adobe’s “Creative AI” platform, which was rolled out after the company mapped out a revenue‑boosting roadmap centered on automating design workflows for its enterprise customers. By tying AI initiatives to measurable outcomes—such as a 30 % reduction in design cycle time—Adobe ensured that executives could justify the investment.

In contrast, many companies launch AI pilots without a firm link to strategy, resulting in siloed projects that never scale. The article urges leaders to embed AI goals within annual budgets and performance reviews, ensuring that the technology remains a priority rather than an experimental side‑project.

2. Start With High‑Impact, Low‑Risk Use Cases

Winners in the AI race tend to focus on “quick‑win” projects that deliver fast, visible benefits. The ZDNet article references a 2022 McKinsey report that identified the top five high‑impact use cases for enterprises:

1. Predictive maintenance in manufacturing
2. Fraud detection in finance
3. Customer segmentation in retail
4. Chatbots for first‑line support in services
5. Demand forecasting in supply‑chain logistics

The article spotlights Shopify’s “Shopify Pulse”—an AI‑driven tool that analyzes merchant data to surface actionable insights on product performance. Within six months of rollout, merchants reported a 15 % lift in conversion rates, proving that targeted, data‑driven AI could quickly justify the initial spend.

The narrative stresses that high‑impact use cases should be chosen based on clear business value, data availability, and a low barrier to entry. By delivering tangible benefits early, organizations can build internal momentum and secure deeper investment for more ambitious projects.

3. Leadership Buy‑In and a Culture of Experimentation

The article underscores the role of senior leadership in championing AI initiatives. Successful companies empower leaders to experiment, tolerate failure, and iterate rapidly. A highlighted case is Bank of America’s “AI‑First” transformation. The bank’s CIO launched a cross‑functional “AI Innovation Lab” that allowed teams to prototype solutions within a risk‑tolerant environment. The lab’s flagship project, an AI‑powered fraud detection system, reduced false positives by 40 % while cutting investigation time by half.

Moreover, the article points to a study by the Harvard Business Review that found companies with a culture of experimentation experience a 50 % higher adoption rate of new AI tools. Cultivating such a culture requires transparent communication about goals, regular metrics dashboards, and a clear escalation path for technical challenges.

4. Data Quality and Governance Are Non‑Negotiables

Across the board, AI projects falter when they are built on poor data. The ZDNet piece cites IBM’s “AI‑Ready Data” framework as a blueprint for ensuring data hygiene, lineage, and privacy compliance. Successful organizations invest in robust data pipelines, master data management (MDM) systems, and AI‑centric governance structures that enforce standards for model training, bias mitigation, and auditability.

The article links to an in‑depth ZDNet guide on “Building an AI‑First Data Architecture” that outlines best practices for integrating data lakes, streaming platforms, and governance tooling. By establishing these foundations early, companies can avoid the costly re‑work that plagues many AI projects.

5. Scaling Through Modular, Cloud‑Native Architectures

Early adopters often use cloud‑native services to accelerate development and reduce operational overhead. The article references Microsoft’s Azure Machine Learning Service as an example of a platform that enables teams to package models into reusable “AI modules” that can be deployed across multiple business units. This modularity allows organizations to scale successful pilots without duplicating effort.

Linked to the article is a ZDNet piece on “AI‑Driven Automation on AWS,” which discusses how AWS SageMaker and Step Functions streamline model deployment and orchestrate complex AI workflows. These resources illustrate that leveraging managed AI services can dramatically lower the barrier to scaling.

6. Continuous Learning and Talent Development

The final section of the article stresses that AI is as much about people as it is about code. Companies that invest in continuous learning—through internal bootcamps, partnerships with universities, and external certifications—can maintain a pipeline of skilled data scientists, ML engineers, and business analysts. The piece highlights NVIDIA’s “Deep Learning Institute” as a leading example of corporate training that has helped thousands of professionals upskill in AI.

The article also suggests adopting a “data‑first” hiring approach, where new talent is evaluated on their ability to extract insights from raw data and translate them into business recommendations. This talent strategy ensures that AI initiatives stay grounded in real‑world business challenges.

What Can Other Companies Do?

1. Define a clear AI strategy tied to core business objectives.
2. Choose high‑impact, low‑risk use cases that demonstrate quick ROI.
3. Secure executive sponsorship and cultivate a culture of experimentation.
4. Invest in data governance and quality from the outset.
5. Leverage cloud‑native AI services for rapid scaling.
6. Prioritize talent development and continuous learning.

The ZDNet article concludes that while AI presents unprecedented opportunities, the path to meaningful success is paved with disciplined strategy, cultural alignment, and relentless focus on data quality. For the 92 % of businesses that remain on the sidelines, the lesson is clear: AI can’t be a technology silo—it must be woven into the fabric of the organization’s operating model.