I like to think that every project goes through at least four stages: conception, creation, implementation, and maintenance. If we take it to an analogy, conception is that sweet instant in which the spark arises and ignites our ideas in a real Eureka! Then, this creation envelops us with enthusiasm and drives us with the energy of being generators of something new. Next, we come to the moment of implementation, and this is where the significant challenge begins: how do we go from that almost divine conception and overflowing creativity to a tangible and useful project?

It is at this point that the truth is revealed and where the “great ideas” are separated from the inventions that really make the difference, because, as in every work of art, it is necessary to distinguish between the purity of the conception and the discipline of the implementation.

It may sound very poetic to say it this way, but we can also transfer it to the technical field. One of those cases, and the one that we will cover in detail in this blog, is the creation and implementation of the new CIAT product: the electronic invoice anomaly detector. Conceived as the perfect complement to electronic invoicing, this project is designed to add value but also requires a solid implementation process. Its design was possible thanks to the support of the CIAT member countries, led by Spain, while the technological development was in charge of Microsoft engineering, within the framework of the agreement we signed for the Center for Advanced Analytics and Artificial Intelligence.

Several key questions arise during this process: who are the actors involved in the implementation? Who benefits from the tool being effective? How is it used? What advantages does it offer from both the planner’s and the executor’s perspective?

At this point we have the experience of the Costa Rican Tax Administration, who were pioneers in implementing this implementation. And we learned something fundamental from that experience: for this solution to be really useful, it is not enough to have the platform working technologically; it is key to understand it comprehensively, both technically and functionally.

This implementation deserves a special mention, as results were not long in coming. Some of the findings today are part of research that is in the legal sphere.

One points that we must highlight is that installing a technological system is just a first step in the implementation process. Accompanying the tax administration throughout the process is essential, contributing not only with technical deployment, but also in a deep understanding of what the tool is capable of doing.

Detecting an “anomaly” is not simply about generating an indicator on a screen: it involves a logical and analytical process in which the information makes sense and allows the tax teams to make strategic decisions. In this sense, the true value of the tool arises when officials understand the reason for each signal, how to interpret it and how to act accordingly.

So, what can we expect from an electronic invoice anomaly detector? The answer is broad. It is a tool capable of identifying suspicious patterns in huge volumes of information, focusing audits on the cases with the highest risk and optimizing valuable resources that are often scarce in tax administrations. But, beyond that, it allows us to generate trust: trust of the teams that see concrete results in their work, and trust of the citizens who perceive that the tax system works with greater equity.

For planners, this tool becomes a compass that points out where the risks are, where it is convenient to pay attention and how to prioritize actions. For the executors, it is a practical companion that facilitates the day-to-day, offering clear information that guides the audits and multiplies the effectiveness of the effort.

And here we return to the starting point: every project starts with a spark, with an idea that ignites the imagination. But what measures the true value is what it manages to transform. The CIAT anomaly detector is a good example of how this spark turns into a real solution, which accompanies administrations, learn from their experiences and adapts to their needs.

Another important aspect is by the tax administrations are not alone in this path. CIAT can accompany them at every stage of the implementation because this is not just about installing software, but a much deeper process.

Accompanying means listening, understanding the context, guiding in the reading of the data, and translating those signals into concrete actions. That is the difference between a tool that “works” and a tool that transforms.

Costa Rica showed us the way, and the challenge now is that more countries feel encouraged to walk through it, get interested in what the tool can do, explore its scope, and discover how that “Eureka!”, this initial impact can be transformed into a tangible impact for tax management.

And as in any project, after the spark, the creation and the implementation, there is one last key link: maintenance. Because projects are neither linear nor static; they evolve, adjust, and strengthen with time. Today, this task of administering, updating, and managing both the data repository and the tool itself falls to CIAT, allowing it to continue supporting its continuous improvement.