Nearshore Americas
cognitive computing

Q&A: The Pathways and Apps Behind a Cognitive Computing Future

Mars Cyrillo brings a unique perspective to his role. He is vice president of products and cognitive solutions at Ci&T, but a background in physics — plus experience in marketing and customer service — makes him one of a kind for the company. His academic work has given him a deep understanding of cognitive computing, including concepts like genetic algorithms, optimization, parallel computing, and machine learning, and he is putting it to use for the firm.
Cyrillo, who moved to his current role at the Campinas, Brazil-based CI&T one year ago this month, talked to Nearshore Americas about cognitive computing, the evolution of machine learning, and what it means for business in the digital age.
Nearshore Americas: Cognitive computing is an approach that draws on machine learning. How is this trend shaping the current business environment?

Mars Cyrillo of CI&T said that all companies will start to adopt cognitive computing in the near future.
Mars Cyrillo of CI&T says that “in one or two years every single company will be looking to build applications that will be using cognitive computing.”
Mars Cyrillo: When I started the product unit at CI&T, one of the areas that we were thinking about using in products was big data and machine learning. I am now completely focused on machine learning and other algorithms that we can call cognitive computing because they are focused on problems that are more human-like.
There are many definitions of cognitive computing, but more and more people in the market understand that it means building systems capable of solving problems that are not very well defined. The variables change over time, so you need to collect a lot of data. You need to collect information about the context of the user and then make decisions based on questions like “what is the best answer to give for this user?” rather than “what is the right answer?” Because there is no right answer. Things evolve, and as things evolve the algorithms might figure out different paths that — while not necessarily the right answers — offer the best solution for that user in that context.
Everything is going cognitive. As the IT industry evolves — and we are talking about being able to reach any person with the right message — you will have to take into consideration that what a person thinks is best for them will change over time very quickly. So there are no rules. Cognitive apps learn about you.
If you have a software that helps you book trips, it could only take what you say into consideration: “I want to travel to Cancun next month and spend a week there.” But it could also look at patterns, such as I always go on vacations to the beach and I always stay in five-star hotels. This could be used to propose things I wouldn’t usually consider. It can be highly motivational for me, because it matches the pattern that the application knows I follow. This is the direction I think every single application will move. Cognitive computing is about getting all the data available and making sense of it.
NSAM: How are you integrating cognitive solutions into CI&T’s product line?
Cyrillo: There is growing interest. We have a customer in Europe with which we are discussing an application that will help them figure out meeting room allocations in a smart way. It is easy to code an application that can help people book rooms, but the end state of that is that everyone will book everything. At any given time you look at all the rooms and they are all booked. So when I need a room suddenly, there is no room available. But if you go and walk around the company, you can see that many of the rooms are not being used.
Part of it will be an application that measures if people are really in the room using sensors, but it will also look at the patterns like Mars always books but never uses the room. That could be one of the insights from the application, and that insight surfaces automatically.
Whenever a customer asks for a solution for a problem that is not possible to solve with simple rules, we try to approach it in a way that collects data — analyzes data — to find patterns and uses machine learning and cognitive computing. We have proposals being discussed in the U.S as well.
The other area is the product area. Our focus at the moment is helping companies really communicate and collaborate — not just to providing another tool for you to post things and chat to employees. What we want to do is provide the infrastructure to bring things together.
We have a new product coming out in March that aims to tackle that problem. It is a “nudge engine,” something that runs in your company behind the scenes to figure things out and then “nudge” people to make them connect.
In one or two years every single company will be looking to build applications that will be using cognitive computing. By then we won’t be talking about cognitive computing because cognitive computing will be computing.
NSAM: What verticals are using such solutions and for what types of applications?
Cyrillo: There are certainly some verticals that are really looking into this. Security is one because it usually uses pattern recognition. The financial industry is ahead of the game and has been for a while.
Since most businesses are becoming digital businesses everyone will have a common challenge. The challenge is: “How do I get in front of my potential customer?” and “How do I provide the best service to them in the digital space?” Everyone will need to build solutions that will use cognitive computing to better service their customers. Some verticals will build their business models around the use of cognitive computing and make that their competitive advantage. But I think everyone will need to use those types of solutions.
Many people are just waiting because there aren’t many standards right now. But new capabilities will emerge. It is just like the brain. As we evolved from lower primates to human beings, there was a new layer on top of our brain — the frontal cortex — that deals with more complex concepts
Companies like Google, Microsoft, and Apple will be building services that will be more complexity in terms of what they can abstract. Google has provided you with a framework where, if you give it an image, the computer will tell you what is in that image. That allows you to build applications that can see things. These can range from very simple applications to complex robots driving about and making decisions based on what they see. In the future there will be more and more applications that will add extra layers in this capability. That’s how applications will become more and more powerful because they will build on top of each other — just like our brains evolved.
NSAM: What types of privacy issues are there concerning software that learns about you from your context?
Cyrillo: There are many problems arising from applications that can connect the dots. You do one thing here, and then a month later you do three or four things there. A person would need to pay very close attention to connect those two dots, but a computer could do that in milliseconds. So at some point we will need to figure out how to look at those things and regulate them. There will be problems in terms of privacy — and other issues — but people don’t care right now. They want value. It is almost like we are in love with those possibilities, and we are saying whatever you need I will give it to you as long as I am happy with my application. But in the future I think people will come back and say you went too far and too fast.
If you want to build really incredible applications you need enormous amounts of data. The companies that are ethical will try not to use personally identifiable information, though.
NSAM: Two of CI&T’s digital predictions for 2016 are around artificial intelligence and virtual reality. How do these feed into the cognitive solutions trend?
Cyrillo: What we wrote about in terms of artificial intelligence is the progression. If those evolutions are not a perfect exponential progress, they are very close to it. They are doubling in capability every year. The number of applications that use machine learning or other types of artificial intelligence will be 10 times greater when compared to 2015.
VR has been there for a while, but it reached an interesting moment when the hardware needed to render VR is becoming extremely inexpensive. So it will be in the hands of millions of people, and then brands will be willing to invest dollars and that will cause an explosion of investments. And machine learning will be part of it. If you are experiencing a virtual world, you cannot code every single thing in there. You need to bring elements that are making sense of the world.
VR right now is not fully using that type of technology, but they will. They will do things that we couldn’t even imagine 10 years ago.
NSAM: How do you see cognitive computing solutions evolving in the future?
Cyrillo: I truly think it will affect every aspect of our lives. The most dramatic change is how the algorithm will evolve with you. You are born and you grow up, and there is an algorithm basically following you. It might follow you through your life, learn about you, and teach you new languages and new skills. You will have virtual friends that were coded to evolve as you evolve.
That will be some sort of entity that is really smart, but not a computer that is smarter than everyone and wants to kill everyone. It will be something more practical. That’s the future I would like to have.
NSAM: How is the nearshore addressing the drive towards cognitive solutions? What needs to be done for Latin America and the Caribbean to leverage the potential of cognitive computing?
Cyrillo: Cognitive computing is not yet being embraced in Latin America. I think Latin America is still having its romantic moment with Big Data, but I am not really seeing products that are cognitive.
The most important thing is to believe! I have been out of Brazil for seven years now, and I think that Brazilians are very skeptical sometimes. I am not speaking for other countries, but it takes two or three years to start doing something. Now they are finally getting Big Data.
We need leaders in Brazil to maybe take advantage of the crisis we are facing there. We don’t have many options, so you have to try something new and different in order to create something really valuable that will outsmart everyone.
It is not for the lack of intellectual power. We have lots of smart people in Brazil. We have lots of investment, too but the focus is on taking lower risks. I am cheering for Brazil, though. I hope they say “we can do it as well, let’s do it faster, let’s do more.” Then you will see incredible things happening.

Sign up for our Nearshore Americas newsletter:

Bianca Wright

Nearshore Americas Contributing Editor Bianca Wright has been published in a variety of magazines and online publications in the UK, the US and South Africa, including Global Telecoms Business,, SA Computer Magazine, M-Business,, Business Start-ups, Cosmopolitan and ComputorEdge. She holds a MPhil degree in Journalism from the University of Stellenbosch and a DPhil in Media Studies from Nelson Mandela Metropolitan University.

Add comment