Digital Transformation Requires Solid Foundations: Xertica.ai

Q: Xertica.ai was born from the fusion of three companies across different countries and cultures. How do you lead an organization that must integrate from day one rather than simply grow?

A: We built Xertica.ai with the explicit ambition of scaling a unified business model across Latin America, and we determined early on that strategic acquisitions were the fastest path. We started with companies in Mexico, Colombia, and the Southern Cone, later expanding into Brazil. The conventional wisdom is that mergers fail on culture, and that is true, but the degree of difficulty is significantly reduced when the business models are closely aligned. Spreadsheets can reconcile quickly; people and culture require a longer, more deliberate process, but a shared operational logic shortens that runway considerably.

Q: What is the most common mistake leaders make when attempting transformation in sectors historically resistant to change, including governments and judicial systems?

A: The most persistent mistake is attempting to solve the most complex problem first. Digital transformation, like any capability-building process, requires sequenced learning: start with contained pilots, build accountability, and scale gradually. COVID-19 accelerated this across both public and private sectors by forcing digitalization at an unprecedented pace. The launch of ChatGPT in 2022 compounded that momentum. Organizations that skipped the foundational steps in pursuit of advanced AI applications are the ones struggling most.

Q: Xertica.ai has deployed AI in contexts where decisions directly affect human lives, including the justice system. What decisions should never be delegated to an algorithm?

A: Our design principle is consistent: AI empowers people; it does not replace them. In justice applications, we accelerate evidence gathering, investigation workflows, and information synthesis, but the judge or prosecutor retains full decision-making authority and accountability. The risk we guard against is misplaced trust: both at the consumer level, where people treat AI outputs as objective truth, and at the enterprise level, where insufficient "grounding" allows models to respond beyond their sanctioned knowledge base. Human judgment, the capacity to distinguish right from wrong in context, must remain non-negotiable at every decision point that carries real-world consequence.

Q: What will make 2026 the year of Agentic AI?

A: Since 2022, most organizations have interacted with AI purely as a conversational interface: asking questions and consuming responses. That is a fraction of AI's actual capability. Agentic AI marks a qualitative shift: instead of querying a system, you commission an agent to execute tasks autonomously over extended periods. A study presented at Davos by ChatGPT quantified this gap directly: individuals leveraging Agentic AI are seven times more productive than those using standard chat interfaces. The tools to deploy agents without deep programming expertise — including Claude and others released since late 2024 — have now crossed the accessibility threshold. That is what makes 2026 a structural inflection point, not just an incremental upgrade.

Q: What misconceptions have you identified in the adoption of AI in Mexico?

A: We surveyed 150 companies — spanning legacy corporates and digital-native firms like Kavak — and found that both groups rated their AI maturity at nearly identical levels. That self-assessment is inaccurate. The disparity between chat-based AI users and Agentic AI users represents a sevenfold productivity gap, yet most respondents do not yet have the conceptual framework to distinguish between the two. They are measuring themselves against a standard they do not fully understand — a classic case of not knowing what they do not know. The transition to agentic workflows is not yet mainstream, but the organizations that close that gap first will operate in a fundamentally different competitive position than those still treating AI as a smarter search engine.

Q: How does Mexico compare regionally and globally in AI adoption, and what would it take to close the gap?

A: Mexico is not punching its weight. Despite having the second-largest GDP in Latin America and a substantial research base, it is not producing AI at the level its resources suggest it should. Countries like Chile, Uruguay, and Brazil are meaningfully ahead. Chile chose to compete on talent through a public-private initiative called "Hazlo por IA," targeting SMEs and government officials, based on findings that 50% of SME workflows could be improved by at least 30% with AI. Brazil bet on infrastructure and now operates world-class AI processing capacity. Mexico has not yet chosen a strategy. Meanwhile, countries like Pakistan and Vietnam lead globally in "thinking token" consumption, the metric that most directly reflects Agentic AI adoption. Mexico's path forward requires deliberate alignment between government policy, public investment, and private sector execution.

Q: Xertica.ai works with governments managing data for millions of citizens. What are the real risks of deploying AI in the public sector?

A: The regulatory challenges are real but solvable. Data protection legislation, judicial confidentiality requirements, anonymization protocols, and encryption standards all exist and can be satisfied with current technology. What actually consumed the most time in public sector digitalization was not AI, it was the transition from on-premise infrastructure to the cloud. That migration required the deepest institutional change. Once cloud adoption was normalized, AI integration followed with considerably less friction. Brazil is an instructive benchmark: 99% of its justice system is now digitalized. The hard part of the compliance and infrastructure journey is largely behind the region's leading markets.

Q: Nearly 10 years after founding Xertica.ai, what still surprises you and what do you believe is still missing?

A: When we founded the company, our core thesis was that a gap existed between where transformative technologies were being developed, primarily in the European Union and North America, and where they were actually being implemented. Cloud computing eliminated the physical latency of that gap almost overnight. AI is now doing the same. What persists, however, is "capability overhang" or the growing distance between what these technologies can deliver and what organizations are actually extracting from them. Our role is to bridge that gap, accounting for the realities of our markets: uneven connectivity, diverse educational baselines, and hardware constraints. The technology increasingly assumes ideal conditions that do not yet exist uniformly across Latin America. Closing that gap is still our goal.

Q: What question should CEOs who measure success exclusively in business metrics be asking themselves?

A: The question should be: how does what your business does affect the society it operates within? Early in any company's life, survival and profitability are the right priorities. But at scale, the outputs change in nature. When you automate judicial processes affecting millions, or conduct a national census, the accountability extends well beyond operational KPIs. 

We made a deliberate choice to anchor Xertica.ai's purpose to four UN Sustainable Development Goals:  Justice and Strong Institutions, Quality Education, Technological Infrastructure, and Health. That is not a marketing posture; it is a strategic filter. The CEOs who will lead the next decade are those who can hold both the business case and the social consequence in the same decision framework, simultaneously.