Over the last few articles, we've explored why it is necessary for organizations to embark on a digital transformation journey. We've explored that the reason is mainly that business development is not moving at the same exponential rate that technology is, and it is necessary for organizations to be able to digitally transform themselves for them to become competitive in this new world. This week, we will be looking at the different intricacies of two main technologies that we've been studying and working on over the last few years, the first being artificial intelligence- AI for short.
For most people, just the mention of artificial intelligence scares the living daylights out of them, mostly because a lot of the different knowledge and information we've read about it has been in a negative light. Everybody remembers that famous movie from the 80s, whereby the robots and the machines were taking over everything that the humans owned, and the robots were taking over and doing things per their will. That is what we call general artificial intelligence (general AI). Unfortunately, or rather, fortunately, pure intelligence hasn't reached its peak, whereby it is able to reach those kinds of knowledge levels, or levels of operation where it can be able to do that.
On the other hand, we've got narrow AI, which is quite simply said, a more specialized variation of AI that specializes in specific fields of expertise and is not necessarily a know-it-all being that is general AI. Narrow AI, on the other hand, seeks to perform a specific function of the human mind. For example, you'll find that the kind of applications that we've been able to build narrow AI on have been solving specific problems, e.g., facial recognition software, and not necessarily multifaceted issues, such as those performed by general AI as depicted in the movies. With that said, what then can this AI phenomenon do for you and your business? It has been theorized multiple times by academics, you're reading about it here today, and from multiple parties around the world talking about how it is the next wave in the world.
Case Study
In typical fashion, we will go on and take a case study of an organization that we recently worked with within the SADC region. This organization is a property development company that specializes in taking houses or commercial properties, buying them up, once they buy them, they improve the property and then go on to resell it at a higher price in the market. What they wanted to do was to find a way to arbitrage the market by finding undervalued properties, improving them, and making a bigger spread on the different properties they sold. So, what we did was, we took a large data set of house prices across the nation from the Deed’s office. Once we took all this large set of data, we then went on to clean it up, removed all the different inconsistencies, and ensured that it was a clean set of data. Once we ensured that was a clean set of data, we then took this data and then created multiple machine learning algorithms, which really is a branch of AI. And with the use of these different algorithms that we created for them; we were then able to find the different factors that affect the pricing of these different houses and were able to find a source of arbitrage that help this organization get a unique position within the market. As a result, they were able to easily spot houses that were underpriced in the market and were now able to make a bigger spread on the different houses they bought and sold. Over the last two years of their using the system, they've managed to make additional profits of up to 75% year-on-year as an organization. This is truly remarkable and illustrates the power of using these new technologies.
Building Blocks
From this particular example, what are the different building blocks that can be found and used in an organization for AI to be implemented successfully?
First, is data - high quality, clean data. It is very important that all the different data sets that organizations need are structured. And the data is clean for us to be able to create custom algorithms that will help us find these different paths to decision making. And once you have the clean data, we can then use different types of models. Within machine learning, which is another branch of artificial intelligence, there are two types of models that we can use. These are:
- Supervised learning - has to do with structured data, which is, data that can be structured such as name, ID number, grade, and so on and so forth.
- Unsupervised learning - is a more complicated version of machine learning, which takes unstructured pieces of data and tries to make sense of them over time, and then goes on to help make decisions around it.
Drawing from the case that we have just illustrated, you can see that the kind of data that we used in this particular exercise, varied from some structured data, including information from the Deeds office, some from unstructured data, such as negotiation price points, some from other data, including neighborhoods, and so on and so forth. And from these different sorts of data that we collected, we're able to then come up with these different models that went on to help organizations.
Benefits
So now then, what are the direct benefits of this artificial intelligence phenomenon? And how will they enable you as an organization to become better, faster, cheaper?
- Finding a unique selling point - they will definitely help an organization find a unique selling point or a competitive advantage that will help them thrive within their market. For example, the above property client was able to find undervalued properties much faster and much more efficiently compared to their peers. Thus, were able to have a competitive advantage.
- Reduce cost base - organizations can use artificial intelligence to help reduce their cost base as an organisation. For example, we recently worked with an organization that was automating its production line. And through the use of artificial intelligence, we're able to reduce the total number of people on the line by over 50%, using artificial intelligence technology. And as a result, we're able to get much more optimized results for a cheaper price than that of individuals.
- New growth opportunities - artificial intelligence has opened up a new avenue for different organizations to find new growth opportunities that they will not have been able to leverage earlier on before it, as this phenomenon uses data to find these different insights that are hidden.
Drawbacks
With the benefits of AI, comes potential drawbacks that we have seen in our client’s experience. For example, in some cases, we realize that artificial intelligence becomes a bit of a black box in that the different results within that black box cannot be thoroughly defined as to how they were derived. For example, a large bank in the United States of America used artificial intelligence - machine learning specifically - to help them decide which individuals would be credit-worthy and which would not. And in doing this, the algorithm took all the different types of data they got, and in a black box then decided that black people would be more likely to default than white people. And at the end of the day, this caused a huge problem because white people that had a similar profile to black people would be getting their loans at a cheaper rate than black people with the exact same profile. And this led to regulatory risk for these individuals as well as reputational risk for them as an organization. As of now, they really didn't know what was really in that model.
The second big drop that we have realized in implementing AI is that it requires a lot of computing power in order for us to execute large tasks. So as a result, this large compute power is going to require large servers as well as large cloud service costs that might be costly for an organization.
The last major drawback that we find from artificial intelligence is that due to its rapid improvement over the years, we find that in the next couple of years, it will completely eradicate and remove people from jobs to an extent. As a result, it's necessary for individuals to look at what jobs are actually in trouble at this particular point in time, and then pivot or rather specialize in different areas that will ensure that they can coexist with artificial intelligence.
However, all in all, the rise of artificial intelligence as a technology has been overall good for organizations across the world. We've seen organizations reduce their costs by up to 70%. In some cases, and with some organizations across the world, they have even gone on to double the size of their businesses using the advent of artificial intelligence to give them the benefit. As a result, in implementing this technology, although there's a large investment behind it, the benefits are likely to come for your organization given that the implementation is right.
Our world is changing. And the truth is, we need to understand what we as organizations must do to be competitive. As organizations are using these different technologies to improve themselves and leverage them to become better, faster, and cheaper, it's necessary that an organization invests in the future for them to ensure their long-term sustainability. Artificial Intelligence, from our experience, and from the insight we’ve drawn from working with other larger organizations is a sure way to be able to do this. However, as an organization before you decide to invest in it, you need to realize and you need to understand that there has to be a business case you are solving in order for this to be successful.
Thank you and please do join us in the next article as we continue the journey to help you and your network ensure that you become better, cheaper, faster.
Tavonga is the Managing Director of Xavier Africa, a software development agency based in Gaborone, Botswana