Rogério Lam, a board member of BCI, is the first guest on the podcast “Conversas Digitais” (Digital Conversations), a joint editorial initiative by Economia & Mercado, Diário Económico, and EY dedicated to the transformative potential of technology.
In a global context marked by the rapid expansion of artificial intelligence (AI), Mozambique is beginning to take steps—still modest but strategic—toward incorporating the technology into key sectors of the economy, such as banking, healthcare, and education.
Despite structural constraints, particularly in terms of digital infrastructure and Internet access, there is an opportunity to accelerate development, provided that technological adoption is implemented intelligently and adapted to the country’s socioeconomic reality.
Bruno Dias, Office Managing Partner at EY in Mozambique, and Rogério Lam, a BCI board member, analyze the influence of AI on the economic landscape, the challenges related to data governance, ethics, and system reliability, as well as the implications for employment and workforce skills.
This episode marks the launch of the EY Digitalks – Digital Conversations Podcast, the first format of its kind developed through a partnership between Media4Development (owner of Revista E&M, Diário Económico, and 360º Mozambique and Angola) and the consulting firm EY. It aims to provide a regular platform for strategic thinking on innovation, digital transformation, and the future of Mozambique’s economy.
Technologies once seen as science fiction are becoming reality, yet this evolution coexists with a country still facing infrastructure deficits. How can these two worlds be reconciled?
Bruno Dias: It is true that Mozambique faces significant structural challenges, from connectivity to technical capacity and the population’s level of digital literacy. However, there is one element that I consider absolutely strategic: demographics. The average age of the Mozambican population is very low—around half of the population is 15 years old or younger. There is a generation that will grow up already integrated into this new technological context.
There are already concrete signs of technological adoption. Some large companies have begun developing pilot projects using artificial intelligence, particularly in data analysis and machine learning models. We ourselves have participated in projects that use these concepts to extract value from information.
There are also examples outside the private sector, such as a pilot project using AI for the early detection of tuberculosis through the analysis of X-rays. These cases show that even with infrastructure limitations, it is possible to begin applying these technologies pragmatically. The most important thing is to understand that the potential for growth is very high.
Rogério Lam: People often think AI emerged with ChatGPT, but what actually happened was its democratization. AI had already been present in our daily lives for several years, but in a less visible way. Whenever someone accesses YouTube, Netflix, or another digital platform, they are interacting with algorithms that learn from user behavior. That is artificial intelligence.
In Mozambique, the main challenge remains access. Although most people have a mobile phone, only a portion of the population has regular access to mobile data and a quality Internet connection. The real impact of AI will only be fully achieved when this access is expanded to the scale of the population.
“In Mozambique, the main challenge continues to be access. Although most people have a mobile phone, only a portion has regular access to mobile data and a quality Internet connection.” — Rogério Lam
Even so, considering the country’s demographic structure and population growth rate, Mozambique has the conditions to make significant technological leaps in the medium term, particularly if there is investment in digital infrastructure and skills development.
In the banking sector, are these technologies already being applied in practice?
Rogério Lam: Absolutely. International banking has been using AI intensively for several years, and in Mozambique we are beginning to see practical applications.
One of the clearest examples is fraud detection. By analyzing behavioral patterns, systems can identify suspicious transactions almost invisibly to the customer but with great effectiveness.
Another important area is predictive models. Today, based on historical information, it is possible to predict the probability of customer default, assess risk profiles, and support lending decisions. These models allow banks to make more informed and consistent decisions.
There is also enormous potential in the use of language models, such as chatbots, for customer communication. Instead of lengthy technical manuals or unintuitive product sheets, AI can explain products and services in natural language, closer to the way people actually communicate. This improves the customer experience and reduces information asymmetries.
Bruno Dias: It is essential to ensure controlled environments. Any information placed on open platforms can get out of control. For this reason, companies create closed environments with their own protected data.
The next major leap will be the transition from large language models to what are known as Large Action Models, which integrate actions, movements, and robotics. That is where we enter a territory of profound transformation.
But this raises issues of control, reliability, and risk. How can we ensure AI does not provide incorrect information?
Rogério Lam: This is one of the biggest challenges today. AI can produce extremely well-structured and convincing answers with excellent linguistic quality, but that does not necessarily mean the information is true. The phenomenon of so-called “hallucinations” is real and requires extra caution. In the banking sector, this risk is particularly sensitive.
If we are using AI to inform customers about products and services, we must ensure the information is accurate, precise, and aligned with the bank’s real offerings. That requires a significant effort in training models and controlling content.
This is where data governance comes in. It is not just about technology, but also about processes. It is essential to know what data enters the systems, its quality, who has access to it, and for what purpose. A model trained with poor-quality data will produce poor-quality results.
“Some large companies have begun developing pilot projects using artificial intelligence, particularly in data analysis.” — Rogério Lam
In addition, decisions must be auditable. A bank cannot have a model that decides to grant or refuse credit without being able to explain to the customer or the regulator how that decision was made. Ethics, transparency, and accountability remain human responsibilities—they cannot be delegated to a machine.
Bruno Dias: Training is central. All employees need to know what they can and cannot do with these tools. Digital literacy and AI literacy vary widely within organizations, and this must be addressed.
In a country like Mozambique, where can AI have a direct impact on daily life?
Rogério Lam: Mozambique has the conditions to benefit significantly precisely because it starts from a lower level of development.
In healthcare, for example, AI can support remote diagnostics. Basic patient data collected in a remote district can be analyzed by machine learning models, allowing initial triage and facilitating the work of doctors located in provincial capitals.
In education, the shortage of teachers and the high number of students per class are well-known challenges. Digital platforms and intelligent assistants can complement in-person teaching, supporting students individually and enabling closer monitoring even in resource-constrained environments.
In agriculture, the impact could be enormous. Through satellite images and AI models, it is possible to map crops, forecast harvests, monitor land use, and support agricultural policy decisions with a level of precision that would be impossible with fieldwork alone. This increases state efficiency and improves sector productivity.
Bruno Dias: We have already carried out projects in which AI combines meteorological and soil data to predict agricultural yields with high accuracy, replacing decisions based solely on empirical experience and significantly increasing productivity.
What about the impact on employment? There are concerns that AI may replace people.
Rogério Lam: That concern is legitimate, but I believe we are mainly facing a transformation of roles. As in previous technological revolutions, some tasks will disappear, but others will emerge.
AI is, above all, a productivity-enhancing tool. The key difference will be the ability to use these tools intelligently. Those who know how to ask good questions, interpret results, and make informed decisions will have a clear advantage. AI is a powerful support tool, but it does not replace critical judgment.
Ultimately, AI forces us to be better.
Rogério Lam: Exactly. It pushes us to think better, question more, and exercise sharper critical thinking. Technology accelerates processes, but responsibility, ethics, and strategic vision remain on our side. The future of AI will always depend on the choices we make as a society.
Watch the video here: https://youtu.be/ManaPKuL60o
Text: Nário Sixpene • Photography: Mariano Silva
