Argentine economist Julian Colombo recalls that, during his time in a bank, the institution operated with 72,000 software to carry out its daily activities. “A software is like a living organism. If I give you 72,000 cats to take care of today, you don’t survive the weekend. So, the solution is not to add more software, but to eliminate it,” says Colombo.
With this idea in mind, after 20 years in the banking sector, Colombo founded N5, a software company specializing in the financial industry. Currently, N5 has a presence in 15 countries and has clients such as BCP, Santander, Credicorp Bank, Fundación Dondé and Zurich Insurance.
The biggest challenge Colombo has identified in the sector is to reduce what he calls “technological entropy.” This concept, inspired by the second law of thermodynamics, describes how technological systems tend to clutter over time, especially when they accumulate multiple legacy softwares that do not communicate well with each other, making them obsolete and complex to manage.
“I realized that thousands of software can be replaced by few, because financial institutions are summarized in five key areas: managing customer information, handling employee data, conducting transactions, orchestrating communication channels and storing information securely,” explains the CEO of N5. “They represent 80% of a bank and can be done with a single software,” he adds.
To meet this challenge, N5 offers a comprehensive platform that integrates solutions such as digital onboarding, customer relationship management (CRM), advanced analytics, and real-time campaigns. In addition, since this year the company began to offer three artificial intelligence (AI) tools: Alfred, Pep and Singular.
Alfred, inspired by Batman’s iconic butler, is a virtual assistant that automates administrative tasks for bank executives, such as managing customer communication. Pep, on the other hand, acts as a digital “coach” who trains agents to provide the best customer service possible. It offers two modalities: “sparring”, to practice and improve skills, and “simulation”, which recreates interactions with potential and real customers using CRM data.
Finally, Singular combines the functionalities of Alfred and Pep to create a “digital executive” capable of performing various functions such as collections, sales, customer service or investment advice. This tool can take on different personalities and tones of voice, depending on the role it plays, with the aim of offering the best result.
“These solutions provide statistical advantages to help banks achieve their four main objectives: having more satisfied, more profitable customers, fewer risks of fraud or non-payment, and lower operating costs,” concludes Colombo, who will participate as a speaker at Simalco’s MTC “AI Transformation: The path for non-technical leaders”, on October 10.
– Does N5 use these tools for its own sales or operations?
“We use Alfred and Pep, but we don’t think we need the third one. N5 develops software for banks and insurance companies, but we don’t use our own tools because we’re neither a bank nor an insurer. It’s like selling cotton harvesters: I wouldn’t use one to take my kids to school.
That said, we use AI intensively. Alfred, for example, participates in all of our meetings: he makes summaries, identifies pending actions, and sends reminders. You also have access to our database. If I am going to visit Ecuador, I ask Alfred for a list of nearby employees and he sends them invitations automatically.
– With the emergence of a technology, many companies tend to focus on it. However, in order to stand out in the industry, you must have had the vision 10 years ago. How do you achieve that perspective?
– There is a book called Superforecasting, which analyzes the experts who make predictions. The authors found that they have a worse outcome than flipping a coin. So they created the “Good Judgment” project that asks many people questions to make their predictions: what is going to be the outcome of the war in Ukraine, how high is the price of oil going to be, who is going to win the Champions League?
The interesting thing is that those who have the best results are not traditional experts, and they have certain characteristics. They have a probabilistic approach, consider various scenarios, and are more moderate in their conclusions. They also work as a team. For example, talking with the director of the port of Valparaíso, in Chile, she explained to me that when the tide is high, ships cannot dock, including those that import sulfuric acid. This affects copper production, therefore the price rises. I would never have imagined this, but if I work as a team and talk to people who know, I can start to estimate the price of copper better.
The best superforecasters are intellectually humble, understand statistics, and don’t emotionally commit to their predictions. If you thought a team was going to win the World Cup, but then they change the coach and you don’t reconsider your prediction, you’re not a good predictor. That would be my recommendation: have these characteristics, read a lot and be in contact with the data.
– Within your organization, are there people who say “this is not possible” when talking about new technologies? I spoke with a cybersecurity expert who told me that CEOs are looking to replace their team with AI, but that it’s not realistic.
– In the 60s someone said that they should eliminate the Nobel Prize because there was nothing left to discover. What that person did not understand was the question of time horizons. We do not have the slightest intellectual capacity to imagine what the world will be like in 600 years.
When the first car came out, many people realized that we would not use horses anymore. But a more sophisticated group realized that it was going to exist at the gas station because the horse drinks water and eats alfafa every so often, so the car will also need to refuel. 99% of people don’t think about the consequences of technologies.
Two weeks before ChatGPT came out, if you asked people, they would have told you it was fiction. There’s a long history of ridiculous phrases about what will never happen. The Wall of China is confidence that the plane will never exist. Saying “this is never going to happen” is very risky.
– How do you manage the organizational culture when you introduce AI?
– We always take a three-stage approach. First I say: “Hey, this exists.” Then: “I wish we could do this.” And finally: “We’re going to do this.” This sequence is less imposing and more pleasant, although slower. You could say, for example: “Starting tomorrow, everyone will use Alfred” or cut the Photoshop license and just leave DALL· E for designers. But we try not to impose these changes in a radical way.
Also, in job interviews, we ask if candidates use AI in their daily lives. If someone says no, we don’t hire them. That stance is contrary to knowledge, it’s like not using Google or not looking for books. He is a person who is deliberately plotting against himself.
At the same time, there are areas that are testing more cutting-edge tools. There is a book called Crossing the Chasm that explains the different profiles that adopt technology. While some wait for it to mature, others are looking for something new, even if it’s in beta or has bugs, because they want to be the first to use it. We consider ourselves an innovative company, more than deep tech. We like to take advantage of what scientists and technologists invent at the frontier and then move quickly, so it is good to have these profiles within the company.
– In your opinion, should business leaders have a deep technological vision, regardless of their industry?
“My opinion is no. I know many leaders who don’t know anything about technology, but they lead people who do know and detect what makes the most sense. However, today technology is a differentiating factor in all industries. If you don’t master it, you’re like a boxer who doesn’t know how to use his left hand. You can be champion, but you lose an advantage.
At the beginning of the last century, the economist Ronald Coase defined that companies exist because it is cheaper to organize production within them than to do it individually. But this is no longer true with technology. Why didn’t Facebook, with thousands of employees, create Instagram, which had only 13 and buy it for $1 billion? Because today, with technology, you don’t need all the advantages of a large company to exist or be successful.
