When we think about how we use artificial intelligence (AI) tools, it’s very likely that—both personally and professionally—our adoption process has been organic, intuitive, and above all, self‑taught.

In recent years, we’ve witnessed a quiet yet highly impactful phenomenon for the future of business: a growing, spontaneous use of AI tools by employees, even before organizations formalize policies, strategies, or investments in AI. This bottom‑up adoption reveals more than it seems. It serves as a reflection of a company’s real data culture and, in many cases, an early indicator of its maturity level.

A gap between leadership perception and operational reality

A recent McKinsey & Company report shows that while surveyed executives and C‑levels estimated that only 4% of their employees were using generative AI for at least 30% of their daily work, in practice that number reaches 13%.

The drive to innovate, complete tasks more efficiently, boost effectiveness or competitiveness—or simply explore new tools out of curiosity—often emerges directly from operational teams. In practice, early adoption usually occurs in very concrete tasks: report automation, writing and information summarization, preliminary data analysis, decision‑making support, or improvements to daily workflows. This kind of improvised usage can signal openness to new technologies, but also an implicit trust in the value of data to enhance experiences.

Now, looking at the role of middle management, especially considering the tension between team‑level initiatives and the C‑level vision. Middle managers often face pressure to deliver short‑term results, maintain operational order, and simultaneously innovate. When that balance is well‑managed, middle management becomes a translator and enabler—turning spontaneous adoption into actionable insights.

In Latin America, a highly diverse region where small businesses, large corporations, informal economies, and cultural and productive heterogeneity coexist, this phenomenon translates into available talent, creativity, adaptability, and often necessity. If organizations pay attention, this enthusiasm can evolve into true competitive advantage

Opportunities in bottom‑up adoption

The initial individual push can become an engine for innovation. Through rapid experimentation, employees test new tools, discover alternative workflows, and streamline repetitive tasks. In other words: they generate value before a formal plan even exists.

Additionally, democratizing innovation can provide valuable clues about real operational needs. Early successes within a team often serve as precedents that can be replicated and scaled across the organization.

Promoting a data culture through a hands‑on approach is essential. When data stops being viewed as a purely technical input and becomes part of the daily routine, the organization begins to internalize its strategic value.

In environments where informality, limited resources, or organizational fragmentation compete with the ambition to transform processes, this type of adoption can become a catalyst for accelerating new technologies.

Risks and vulnerabilities—why data governance matters

Like any growth opportunity, bottom‑up adoption also comes with risks. If each team adopts tools independently, information silos may emerge, formats may become inconsistent, and quality control can be lacking. While it’s not catastrophic, it can reduce reliability and hinder the ability to scale.

This is where data governance becomes critical. It establishes clear policies, processes, and responsibilities to ensure data is accurate, secure, and used ethically. Without governance, decisions made without traceability or analytical support—AI‑driven decisions made “off the books”—can lack transparency, oversight, and clear accountability or ethical criteria.

A lack of oversight regarding what data is used, how it's stored, or how it's shared can expose companies to legal or reputational risks—an important consideration in Latin American markets with developing regulatory frameworks.

And just as important: spontaneous usage does not guarantee sustainability. What works in one team, in a one‑off or improvised manner, may not hold up as part of a broader expansion or growth strategy.

Turning enthusiasm into strategy

As a first recommendation, companies should begin by mapping the real use of AI and data across the organization: which tools are being used, in which areas, how frequently, what data is involved, for what purposes, and with what outcomes.

From there, organizations can design a data and AI governance framework based on quality standards, traceability, privacy, access, roles, and responsibilities. Giving this process proper space is essential to establishing the foundation for structured, responsible usage.

Training can no longer be purely technical—it must include ethical awareness, understanding the strategic value of data, the ability to interpret results, and how to make decisions based on information. In the region, analytical capacity is often one of the most valuable assets. In fact, according to a recent SAP Latin America report, 43% of companies in the region already have a positive outlook on AI and plan to increase investment.

It’s no coincidence that a company like Mercado Libre, after exploring real generative AI use cases, is now building a more integrated data and AI strategy. The goal: anticipating financial scenarios, increasing audit productivity, and improving internal talent experiences—a concrete, strategic use of data.

Bottom‑up adoption, in that sense, delivers a clear lesson: organizations must guarantee privacy, integrity, and traceability. And this requires shared responsibility across IT, business areas, and users themselves, guided by clearly defined rules, roles, and a unified data semantics. This federated governance approach enables AI adoption at scale without losing control, aligning operational autonomy with consistency and accountability.

Recognizing data as an essential input is not a minor gesture—it’s the foundation for better tools, better analysis, better forecasting, and ultimately, a key driver of innovation.

In Latin America, the data infrastructure must be flexible, representative, and scalable. That’s why bottom‑up adoption isn’t a sign of disorder—it’s a sign of need. Incorporating intelligent data management into strategic decisions can uncover new products, services, or process improvements.

At a time when digitalization coexists with structural inequalities, having a data infrastructure capable of integrating information from diverse sources, formats, and systems becomes an opportunity to build the future. A conscious, responsible, context‑aligned data strategy strengthens resilience, competitiveness, and the ability to innovate.