Artificial intelligence is evolving from a technological promise into a real tool for operational transformation. In the service sector, where efficiency, data management, and process optimization are key to staying competitive, AI opens up new possibilities for automating tasks, improving decision-making, and freeing teams from repetitive work.

We spoke with Pablo Menor, co-founder of Zelvion Tech, about how to apply artificial intelligence to business processes in a practical, measurable, and results-oriented way.

 

Many companies in the service sector are exploring AI, but they don't always know where to start. What is usually the first business process where it makes the most sense to apply artificial intelligence?

I believe the best place to start is usually with the company’s operational and administrative functions. These are processes that involve a large volume of repetitive tasks, a great deal of document management, and a lot of time spent on activities that don’t really add any unique value.

That’s where artificial intelligence typically delivers a very quick return on investment. We’re talking about automating document classification, extracting information from documents, generating automated responses, assisting employees with internal tasks, or improving customer service through intelligent assistants.

In addition, these are typically relatively small-scale projects with a reasonable investment, and their impact can be measured from the outset. Our recommendation is always to start with specific use cases that solve a real problem and demonstrate value before tackling larger initiatives.

 

Based on your experience, what distinguishes an AI project that truly makes an operational impact from one that remains just a pilot or just another tool?

In recent years, we have seen many companies invest in artificial intelligence projects simply because it was a trendy technology. Many pilot projects and proof-of-concept initiatives have been developed that never made it to production because the goal was not to solve a business problem, but rather to incorporate AI.

A good project begins by clearly defining the problem, establishing clear metrics to measure its impact, and understanding how the solution will be integrated into the company's day-to-day operations.

We always try to validate the solution using a prototype that allows us to measure results from the very beginning. From there, we iterate together with the client until we build a tool that people will actually use.

And something we consider just as important is that not everything requires artificial intelligence. In fact, on several occasions we have recommended that clients develop a traditional solution because it was the most efficient option for meeting their needs.

 

In sectors that are highly labor-intensive and involve repetitive tasks and information management, what types of processes do you see as best suited for automation or AI-assisted tasks?

When we talk about artificial intelligence applied to software—as opposed to robotics—the greatest potential lies in all those processes that involve a heavy administrative and information management workload.

AI can handle tasks such as reading documentation, generating quotes, reviewing contracts, processing invoices, sorting emails, preparing reports, extracting data from forms, or assisting employees in finding information.

The goal is not to replace people's jobs, but to eliminate the most repetitive tasks so that they can devote their time to activities with higher added value—where they can truly contribute their experience, judgment, and decision-making skills.

In our experience, the greater the volume of information a company processes each day, the greater the potential for automation tends to be.

 

You mention the concept of an “operational digital twin.” How can it help a service company better understand its operations, identify inefficiencies, and make better decisions?

Many companies want to incorporate artificial intelligence, but they still haven't solved the problem of data, which is the most important issue.

The concept of an operational digital twin involves creating a digital representation of how the company actually operates. In other words, it involves connecting processes, tools, communications, and data to provide a comprehensive view of operations.

Once that foundation is in place, it becomes possible to monitor processes, identify bottlenecks, simulate scenarios, and make decisions based on objective information. Furthermore, that environment becomes the ideal starting point for effectively incorporating artificial intelligence.

Ultimately, the quality of any AI system depends directly on the quality of the data it processes. Without a solid database and integration between systems, it is very difficult to obtain reliable results.

 

One of the biggest challenges is connecting data, tools, and teams that often operate in isolation. What role does data infrastructure play before implementing AI solutions?

The most important thing of all. Before we talk about artificial intelligence, we need to talk about data.

In many companies, information is scattered across different applications, Excel spreadsheets, emails, or systems that aren't even connected to one another. With information fragmented in this way, any AI solution faces significant limitations.

A very important part of our work involves, first and foremost, integrating tools, consolidating information, and building a robust data infrastructure upon which intelligent solutions can then be developed.

Artificial intelligence is the intelligence layer, but the data infrastructure is what supports the entire system. Without that foundation, it is very difficult to generate real value.

 

Looking ahead to the coming years, how do you think the relationship between AI and business processes will evolve? Will we see companies that are increasingly automated, more predictive, or outright more autonomous in certain areas?

We are already seeing companies that are increasingly automated, where data analysis is crucial for decision-making. We are moving toward a point where AI will not only assist with analysis but also directly influence decision-making.

Any company would want a system that allows it to take action before problems arise.