“Stop what you’re doing right now! There’s a better way to do it…”

It’s close to 9 am, and a crowd of tech-savvy individuals, mostly wearing black and purple hoodies, are queuing up to access the conference room at the Hennessy Park Hotel in the Cybercity of Ebène, Mauritius. A few have braved the cold breeze of the rare winter months on the tropical island, proudly sporting black t-shirts emblazoned with ‘Techverse 2024’. The joy, excitement, enthusiasm and smiles on the faces of the tech community as they try to secure the best seats evoke scenes reminiscent of events hosted by Google’s CEO Sundar Pichai or Meta’s founder Mark Zuckerberg. The presence of these tech giants seems almost palpable.

Mantra Connexions has transformed the classical business conference room, infusing it with a Silicon Valley tech-like ambiance. The master of ceremony, casually dressed, and wearing a pair of brand-new white sport shoes, steps up. “Welcome to Techverse 2024!” The young lady’s American accent, despite being born and bred in Mauritius, silences the audience and captures their attention as the workshop on the next big thing in the tech community, Gen AI, begins.

Accenture Mauritius, in gathering the software and coding community, has clearly succeeded in the challenge presented to it. The first panel discussion, “Future in Tech”, which kick-starts the show, brings together professionals from the private sector and public institutions, including Nicholas Mckenzie, Engineering Lead at MCB; Didier Béphage, Technology Director at SD Worx; Dr. Maleika Mamode Khan, Dean of Faculty at the University of Mauritius; Dr. Riyad Ameerudden, Associate Director at Accenture; and Vrigesh Futta, Executive Director at the Mauritius Emerging Technology Council. Moderating the discussion is Rinayah Appadoo, Senior Manager, Industry Portfolio Lead, and SAP Practice Lead at Accenture Mauritius.

BIZWEEK brings you the main features of that panel discussion.

Introduction by Rinayah Appadoo, the moderator of the panel discussion – When we talk about technology today, we talk about technology which is human-like. Technology that will enhance the potential, the productivity and the creativity of human beings. Before, we used to adapt to technology. We had to change our habits to use an application or an interface or a tool. Today, we’re having technology that is adapting to us. We have to say we’re a painter. I’m not a painter, but I’ve got a creative mind. Sometimes, when I want to make a painting, it doesn’t come right as I would want it to be. Today we have tools. I can tell the tool, I give it the mood, I give it the colours, I give it the paintbrush, and it comes with the painting that I want. It’s all about superpowers. It’s about augmenting our capabilities. That’s what technology is about. It’s not about replacing humans. It’s really giving us superpowers and additional powers to make our life more creative and more fun. This is why we’re here, as fun people in technology. We’ve seen in the tech vision all the different trends. We know this is coming in the near future. 

“It’s also freeing up humans to do things that humans are great at”

Moderator – I have my first question. Nick, you’re going to be the first one here. How do you see AI impacting customer experience and bringing operational efficiency in your sector?

Nicholas Mckenzie, Engineering Lead, MCB – It’s hard to think of an aspect of what we do that is not impacted by AI, which is why at MCB, we’ve put together a multidisciplinary team looking at this across not just people in tech, but across all aspects. People are bringing their experiences, their needs, and what they want to this. We’ve been running this for months now, in terms of trying to understand how this is going to impact us, how we need to change, what the opportunities are. Some of the things are obvious in terms of how it will impact us and our customers in terms of looking at things like chatbots and assistants. We’ve tried to model on some of our frequently asked questions. We’ve connected it to our help desk and we’ve been able to get this thing to actually record and interact with our customers, ask questions, and automatically queue up actions if needed. This is what I think everyone thinks about when we’re just looking at the impact on customers. 

But, it goes much deeper than that when we look at general AI in terms of things like generating completely radical approaches to new products and services, generating analyses of a customer’s financial position, and recommendations that are tailored way, way beyond maybe what would be economically viable for us to do on an individual basis. 

There’s democratisation of expertise that comes through, and we see opportunities for this all across the organisation. All of these things generate operational efficiencies because it’s also freeing up humans to do things that humans are great at, and allows us to then focus on things that AI maybe can’t help us with. 

In terms of pure automation, when you look at what we’re doing around document classification, extracting information from documents, automating workflows, optimising workflows, all of this has an operational impact, and this is before we even get into the software engineering space. The things that we’re looking at are like doing real-time threat detection, training a model that looks at our traffic so we can detect things that are maybe not normal and require attention.

Something else we’re looking at is using AI to help us find what we’re calling the needle in a stack of needles when it comes to analysing our logs. We generate tens of millions of log events daily, and in that huge amount of data, sometimes, there’s that needle we need to see. What we’ve been looking at is using an AI to actually summarise this down to find patterns, to find commonalities. All of this is stuff that would be hugely time-consuming, but something that, with AI, will be much simpler and have a big impact on us.

I believe this will really enhance the customer experiences. I can imagine that, at any time, I can just have a look at my data. I don’t need to go to the bank, request a long sheet, and aggregate all the different accounts and all the operations that I had. Even the things that are not necessarily immediately customer-facing bring efficiencies there and make the teams faster and more responsive, and that has this kind of knock-on effect all the way through the organisation. It’s from top to down.  

“There’s a bit of control needed”

Moderator – Today we have an intersection of technologies coming in. We’re talking about spatial computing. When we talk about metaverse, it’s blending the digital and the physical space. We’re talking about blockchain as well, and in blockchain, we talk about human values, transparency, and trust. So how do you see this intersection in your industry?

Didier Béphage, Technology Director, SD Worx – At SD Worx, we’re a bit of a software success company, so we build our own software. For anybody who’s in a software development industry, the first intersection that you see is that you have a lot of intellectual property that you build, and that traditionally was made by humans, using human creativity. It was tested and peer-reviewed by other humans, and it was deployed by other humans, so you had a long human chain of collaboration. Today, this is getting more and more automated. We had the rollout of DevOps, CSP pipelines and all those things, and that was before AI. What we can see now is that in all the personas that we use in our software development environment, there’s some aspect that can be augmented or automated.

If you take a developer and you need to generate some code, you have a couple of options. You can write it completely from scratch, and I don’t know how many still do that. You go from first line of code to last line of code, you commit, and the code works perfectly. That’s how we learned to do it, and that’s less and less the norm. 

The younger generation in the workforce, they don’t actually know how to do this. They didn’t learn to do it. What they did, in the previous generation, just before let’s say 2023, they would go and stack overflow, copy-paste code, try to run it, find out it doesn’t really run exactly like they want to, and tweak it. We’ve all been programming manually. We know the constructs, the infrastructure, the algorithms, the AI engines are still not spitting proper code that is correct, functional sometimes, and secure.

What they do now is even better. They ask one of the large language models to come up with a code, and it will run. There’s quite a bit of AI in the younger generation, using AI to do their work, but the quality is not exactly there. So, we see higher quality code from the old school developers who don’t use AI, and from the younger developers who rely too much on AI. 

There’s a balance needed here, and this is what we’re trying to address with education programs, and awareness, to use AI correctly and responsibly. 

When you create code and it doesn’t work, it doesn’t execute a function, who’s responsible for it? You, as a developer, and then the tester who didn’t see it. So there’s a joint responsibility. If the AI generated 99% of the code, and the only thing you did was to add the command in your name, who owns the fact that the code didn’t work? It created a massive production, and maybe mispaid an employee, who then lost his house because the man seized it. Who owns that? These kinds of questions are that we need the answer to. 

There’s quite a bit of control needed. Policies, procedures, we don’t like these words, but they are needed to save all these outcomes because these are things that can happen in the manual building world. They can happen even faster if you have a hundred young engineers fully reliant on automated engines, and they are just basically adding their name at the end. 

“We are evolving into new ways of assessing students”

Moderator – Maleika, how do you prepare your students to not directly go into those tools and really make the effort to learn about these principles, which is what we call traditional knowledge?

Dr Maleika Mamode Khan, Dean of Faculty, University of Mauritius – Our role in the industry is to prepare the student, the human capital. We play an important role in that. There are different skills that we have to instil into students. We have the technical skills, the soft skills, the continuous learning, etc. At times, when we talk about new technology, people just think you have the AI, you have the machine learning, you have the blockchain, and we tend to forget about the basics, the principles, the foundation and the baseline. In order to jump into the new technology, we have to understand all the different principles, that is coding, database, architecture, etc. All these basic principles should be there. At the university, we try to incorporate this in our program, in the curriculum, in order to build the different competences that are required. 

Moderator – Are you challenged by your students?

At times, it depends. We have different levels and types of students at the university. Some are very fast and they do challenge us. We are very happy because we try to develop these skills with the students. They need to have a critical mind and critical thinking. We have a group of students who, at times, will challenge us and, of course, there are many innovative solutions that come out of it. 

Moderator – Is it easy to differentiate something that has been generated? You know, I believe that now, when you have to submit a project or report… I think everybody knows what I have in mind…

This is a very big challenge. All of you are, of course, using generative AI. At the university, we have “Turnitin”, and we have one component that does detect what has been generated by AI.  We also have paraphrasing tools. So, even if you can ask ChatGPT to paraphrase some things so that it is not being detected by AI-generated software, we detect it. 

We are evolving new ways of assessing students through different ways of analyzing the data, different trends… We’re basically adapting, as well as using the technologies. 

“It’s almost a change of culture and mind-set”

Moderator – When it comes to all those super powers, super technologies, we talk about our responsibility to be sustainable as well. How do you make your team deliver sustainable products?

Dr Riyad Ameerudden, Associate Director, Accenture Mauritius – Let’s understand that notion of being sustainable, because there is an interesting paradox here. We are building powerful tools. We are having data centres, and the tools, the possibilities that AI and other technologies are using are just amazing, but all this requires a lot of power. We also, then, have the responsibility to be sustainable.

At Accenture, we are a technology consulting company, but at the same time, sustainability is in our mission statement as well. How do we think responsibly? I believe it’s almost a change of culture, of mind-set, of how we understand the impact that those kinds of tools are having on our environment. 

Accenture has set a vision of having net zero emissions by 2025, and it’s next year. It’s quite close and it’s quite ambitious. What is meant by net zero? As a first step, the aim is to try to at least achieve using renewable energy as a source of energy, providing the technology around us, providing our buildings, our offices, and even helping our clients as well. In terms of data centres as well.

I think craftsmanship also includes that notion of sustainability. It’s very important for us to understand that. And that’s why we are pushing our people as well. In every aspect of the design, we try to really infuse that sustainability aspect, be it the hardware or the software, or even in the way we do our coding. 

“A national AI policy to be ready in four to five months”

• We will be very likely impacted by the big decisions of tech companies

Moderator – Mauritius has been ranked the first country in the sub-Saharan region in Africa to be AI-ready. Where are we heading? 

Vrigesh Futta, Executive Director, Mauritius Emerging Technology Council – I think it’s a great thing to be number one in sub-Saharan Africa in terms of AI readiness. We’ve got numerous advantages. I was not in the country for numerous years. I was in Africa, and I’ve travelled across all countries. I then realized how good we are. The first thing that came to mind is that we often criticize ourselves. We often have this habit of saying we aren’t good enough. I think that’s exactly the attitude that has put us where we are right now. So that report was done by Oxfords Insight, and it was interesting because the first thing they looked at was government structure. 

And then you would also say, why would that be? And they would think, okay, how old is the structure? Are you looking into new technologies? Do you have a strategy in place? Do you have a policy? All those questions were asked and measured, and we were the first country to actually come up with an AI strategy, in 2018, which is great work by some fellow friends of mine. 

Afterwards, one of the recommendations of the strategy was to set up the Mauritius Emergency Technologies Council. And it has been set up. 

Our next step, now, is to actually come up with a national AI policy, which we are busy drafting right now. It is expected to be ready in four to five months. We want to get better at operating. We will come up with a policy soon, and it will be very precise in terms of different plans and impact on different sectors of the economy. This will be the next step, obviously. 

I was listening to you talk about sustainability and AI. A good friend of mine, who was sitting in Silicon Valley a couple of months back, told me that every conversation you have with ChatGPT is seven glasses of water. Not in a prompt, but just a conversation that you have. Do you want to be more responsible? Stop doing useless conversations. The computing power generated requires a huge amount of water to cool down the servers. That’s why those at Microsoft put their data centres on seawater cooling. 

Sustainability and AI do not fully go hand in hand right now. Remember that while we talk about the potential, we talk about all the benefits and everything, we have yet to solve the sustainability issue around AI. So I think that’s important.

Coming back to Mauritius, we are, as a country, dependent on large tech companies. We won’t build a data centre here. We need to be humble in our objectives as well. We are not a European superpower. We are not an African superpower like South Africa or Kenya. We will be very likely impacted by the big decisions of tech companies, like Accenture or Microsoft. We need to work closely with them, which we all need. We need to understand what they will release in the country, how we will use that, and the bottom line, the idea right now is, from our level, to ensure that the tools are made available to the companies, to the public at large, to use it. We are talking about augmenting the quantities of humans, companies, and institutions across the board. This is what we are looking into. It’s all about promotion and awareness. It’s all about industry partnerships. It’s all about really telling people to stop doing what you’re doing right now. There’s a better way to do it. 

Just hold on, do some research, and find the right tools to make you get better at what you’re doing.