South Africa Artificial Intelligence Market: Growth Accelerates, Delivery Frictions Persist

What you can see is that companies are really focusing on concentration. Financial services like banks are using intelligence systems a lot and they are spending most of their money on things like finding fraud deciding who gets credit and automating how they talk to customers.

Big banks are saying that more than half of the work they do every day now uses machine learning or automated decision making in some way.

Smaller banks are not doing well with this.

The reason for this difference is that it is the way things are set up. Financial services, like banks are using machine learning and decision automation a lot and this is making it harder for smaller institutions to keep up with financial services.

Standard Bank says they have saved around ZAR 1.1 billion because of their advanced fraud detection models. You see this number a lot. It is hard to check if it is true on your own and Standard Bank does not give details about how they saved from stopping fraud versus how much they saved from being more efficient. Anyway other banks now think these systems are necessary for the banking sector I mean for Standard Bank and other banks, like Standard Bank these advanced fraud detection models are now the norm.

The government has been working on its policies at the time but they do not always match up. The governments National AI Policy Framework is really about using Artificial Intelligence in a way helping people learn new skills and using it in the public sector.

The best example of this is SARS. They are now using Artificial Intelligence and machine learning models to fill out tax returns for millions of people using information from places. In the year more than five million taxpayers got their tax assessments done automatically and many of them even got their refunds in just a few days.

The Artificial Intelligence system is working well. The National AI Policy Framework and Artificial Intelligence are really making a difference in the sector and SARS is a great example of this. Artificial Intelligence is helping to make things easier, for people. This system also brings all the decision making to one place in ways that people do not fully understand yet. The decision-making process is centralised in ways that're not clear, to people. The centralisation of decision-making is something that people are still trying to figure out.

Outside of finance and getting money things are not moving fast. Healthcare and agriculture and manufacturing are still not working together. There are some test projects.. These projects are not being used on a big scale. We still have problems, with getting power when we need it and having computers that work all the time. This affects what we can really do. Some companies we talked to said they are waiting to use Artificial Intelligence widely until they have better energy solutions at their own sites or a mix of different energy solutions. Now having panels is something that companies need to have before they can use Artificial Intelligence.

Skills are still a problem. There are not machine learning engineers, data scientists and AI governance specialists to meet the demand. Many organizations have started training programs. They are working with partners from other countries. For example Microsoft South Africa wants to teach one million people about AI and digital skills by 2026. They are working with non-profit groups and online platforms to do this. This is a big goal. However we do not know how many people are actually finishing these training programs. Machine learning engineers, data scientists and AI governance specialists are still, in supply.

Universities and private providers are making their programmes in data analytics and responsible AI bigger. They think that getting people to learn skills is the answer to not having enough workers.. This is only part of the solution. A lot of companies say that it is still hard to keep people who're good at what they do especially when it comes to data analytics and responsible AI. This is because more and more people, around the world want to hire these people and it is getting easier for them to work from anywhere so they can easily leave their jobs.

The cloud infrastructure investment is what makes people feel hopeful about the future. Because we have space in our data centers and better connections things work faster and we can use things in real time. The government and private companies are working together to make it cheaper for companies and research places to use computers.. It is still not easy for people to get access to these things in many places except, in Gauteng and the Western Cape. Cloud infrastructure investment is really important for this to work.

The story of market segmentation is one we know well. When it comes to intelligence systems the narrow ones are in charge. General artificial intelligence is still an idea. Most of the value is in the software. Then comes the services. The market for hardware is getting bigger. This is because of special accelerators and edge deployments. But there is a problem the supply chains for hardware are affected by changes in currency values. That is a worry, for the market of artificial intelligence systems specifically the hardware part of artificial intelligence systems.

If we look at the regions Gauteng is the one that is ahead. KwaZulu-Natal and the Western Cape are the ones. The other provinces are only seen a bit in the test numbers. People know that things are not fair. This unfairness is still a problem with Gauteng and the other provinces, like KwaZulu-Natal and the Western Cape.

People are using Artificial Intelligence more and more these days. Most adults, seven out of ten are now using Artificial Intelligence tools for their job or to learn new things, which is a big increase since 2023. It seems that people are getting better at understanding Artificial Intelligence than the government is making rules, about it. A lot of companies are saying that they want to follow the POPIA rules. It is hard to find examples of what happens when they do not. The Artificial Intelligence is becoming a part of our lives and we need to make sure we are using it correctly.

The people in charge of companies think they will spend as much money on Artificial Intelligence as they do now. This is often because the CEO wants it to happen. The main reasons for this are to make the company work better and to make the customers happy. The people in charge like to use tools that are just for their type of business especially tools that can understand many languages. The results of using Artificial Intelligence are different, for each company.

The market is growing. Delivery is conditional. Infrastructure, skills, and power remain unresolved. The numbers assume they will not be binding constraints.