The State of AI Adoption in Business: Opportunities, Challenges, and Next Steps

Artificial intelligence (AI) has gone from being a laboratory experiment to occupying a central place on the agendas of companies, governments, and economies. Investment is skyrocketing, adoption is growing rapidly, and more and more studies confirm its potential to increase productivity. However, behind this enthusiasm lies a paradox: although AI used extensively, most companies are still not seeing clear returns. To understand this, we need to look at both the forces driving AI the barriers that are slowing down its real impact on organizations.

The big picture: AI a cross-cutting technology

Narayanan and Kapoor (AI-as-Normal-Technology) describe AI a "normal technology," comparable to other major revolutions such as electricity or the internet. These are general-purpose technologies that change the way we work, but whose adoption is often uneven and slow. Their message is clear: we should not expect immediate and magical changes, nor should we minimize their importance. Adoption is a process of organizational learning, process redesign, and ecosystem building.

The AI Index Report 2025 from Stanford provides data that reinforces this vision. In 2024, 78% of organizations were already using AI (up from 55% in 2023), and global private investment reached 252 billion dollars. Furthermore, governments are regulating and investing on a large scale, and models are becoming more affordable and improving each year. In other words, AI is no longer a promise: it is becoming part of daily business operations, just as electricity once did. But adoption does not automatically mean transformation.

The business reality: the "GenAI gap"

The report State of AI in Business 2025 provides a more stark counterpoint. Despite the $30–40 billion invested in AI , 95% of companies do not see measurable returns. Only 5% achieve integrations that generate millions in real value. The rest are stuck in what the authors call the "GenAI gap": lots of adoption, little transformation.

The causes are not technical. The models are good and the infrastructure exists. What is lacking is integration into workflows, learning capabilities, and adaptation to context. While employees enthusiastically use tools such as ChatGPT for quick tasks, enterprise solutions are often perceived as rigid, impractical, and disconnected from operational reality.

Actual impact: small, but present

Even so, the first benefits are beginning to appear. The AI Index Report 2025 shows savings in operations (49% of companies using AI), supply chain (43%), and software development (41%), although most are less than 10%. In terms of revenue, the most common increases are less than 5%.

For its part, the report State of AI in Business 2025 points out that the greatest returns are not in the most visible areas (such as sales or marketing), but in back-office automation: eliminating BPO contracts, reducing expenses on external agencies, or automating administrative tasks. In these cases, the savings are tangible and rapid, although they do not always involve reducing internal staff.

In more advanced sectors, the impact is much greater. In healthcare, the FDA approved 223 medical devices with AI 2023, and studies show that AI outperforms doctors in certain diagnostic tasks. In transportation, companies such as Waymo and Baidu already offer hundreds of thousands of autonomous trips per week. These examples prove that transformation is possible, but it depends on deep integration into processes.

What companies should do now

Given this scenario, there are several key actions that companies can take to maximize the return on their AI investments:

1. Going beyond pilots: Experimentation is not enough. AI be integrated into processes and redesign them, with systems that learn, remember, and adapt.
2. Look for value in the least obvious places: Most of the budget goes to sales and marketing, but the biggest ROI is often found in operations, finance, and customer support. That's where clear and sustainable savings can be achieved.
3. Focus on strategic alliances: Projects developed with specialized external partners are twice as likely to succeed as internal ones. Companies should negotiate as if they were contracting a process service, not just a software license.
4. Prepare for a gradual talent transformation: AI is not causing mass layoffs, but it is changing the way we work. The key is to equip teams with digital and AI skills, enabling them to integrate into technology-augmented processes.

Conclusion: from adoption to financial impact

The adoption of AI two sides. On the one hand, it is an increasingly prevalent technology, with record investments and success stories in key industries. On the other hand, most companies have yet to translate its use into significant financial impact.

The key to unlocking value lies in stopping treating AI an accessory and starting to see it as an integral part of the organization. Companies that redesign processes, seek value in less glamorous functions, and build systems that learn and evolve will be the ones that take advantage of this new stage.

In short, the benefits of AI great, but only those who are willing to reorganize, relearn, and rethink how work is done will reap them.