top of page
Search

These 3 African Startups are using AI to promote financial inclusion and re-build traditional industries. Find out how...



Africa is increasingly seen as an important startup hub of the world, being, as an example, the only region to experience year-on-year growth in startup funding in 2022. Despite AI being the most fashionable startup trend at the moment, Africa is frequently absent from discussions, which only mention the US, Europe and Asia. However, reality shows that Africa is in fact fully embracing what may be the next tech frontier. It is then only sensible for us to see what developments are taking place in the AI startup space in Africa. 



Written by: Filip Vrábel

-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------



DataProphet



We have all heard of Software-as-a-Service, Platform-as-a-Service or Infrastructure-as-a-Service. But what about AI-as-a-Service? This is the concept driving DataProphet, a South African AI startup. It aims to transform plants into ‘smart factories’, i.e. establishments whose performance is optimised by AI-provided prescriptions which are developed from the client’s data stream which is made “contextualized, current [and securely managed]” by DataProphet. Where might such a solution be particularly useful? Consider foundries: as metal is melted and poured into specially shaped containers, there is a risk that due to the metal’s exposure to the atmosphere (as happens with iron, but there can be various other reasons for problems in the iron casting process and for other metals) there will inevitably be some defects resulting in a foundry’s given scrap rate, ie the proportion of unusable product. AI can be used to detect and correct process anomalies, thus reducing scrap and increasing casting quality. Foundries can engage DataProphet for such a solution because foundries possess “large volumes of production data” (which is sensible considering how much goes through a foundry production line). Moreover, as regards the US or mutinationals, DataProphet offers a cost advantage due to operating out of South Africa. 


DataProphet’s business model illustrates how AI can enhance old industries, especially the secondary sector. Manufacturing is, after all, only a series of steps. Raw materials are taken as inputs and go through an order of unit processes before being shipped as finished goods. This entire process is governed by instructions and parameters (f.e. what temperature should be used for iron melting) so that employees have to be hired to keep everything in check. As we have seen above, if something is out of shape, it is scrapped (or reworked). Since employees can make mistakes, companies also rely on a software component for evaluation and even changing the parameters. The effort is put in to make sure that whatever comes out in the end - the output product - passes a quality check. Mistakes lead to lower yields. DataProphet is a great example of how an end-to-end prescriptive AI can come into this space and be of great utility. The AI can offer advice (ie provide its take on the recipe of instructions the process should follow) so as to “reduce defects, scrap, or non-quality processes and improve manufacturers’ yield” - this is DataProphet’s PRESCRIBE product. What about the contextualization and enhancement of security of data that we have mentioned above? That is the CONNECT product, so named because it enables centralization and collation of data - bringing data from where it has been used for compliance to somewhere it can be used for optimization. 


DataProphet’s competitiveness stems from the cost advantage allowing an international client portfolio - from its home of South Africa to East Asia, Europe, India, the US and Latin America - but also from the very nature of these products. Since this is an end-to-end prescriptive solution, it can look at the lowest data levels, differentiating it from competitors. Another such distinction is that it does not require of its enterprise customers to employ data science experts as the “AI-as-a-service platform [...] thrives on organizing [the] data infrastructure itself”. With the right mix of innovatiness and cost advantage, African AI startups can tap into international sales, as DataProphet illustrates. However, that may not be the sole end of such companies, as the startup below illustrates…




Nomba (formerly Kudi)



Nomba is a Nigerian FinTech startup, which began as "Kudi.ai" - a chatbot integration that “responds to financial requests on social apps”. Today, it is relied on by millions in Nigeria for money processing - as it used to be common for employers to only pay in cash as employees lacked bank accounts. The app has thus increased the number of people even using financial services at all; with the banks not being accessible to everyone, the result is a banking app capitalising on the informal market by digitising formerly cash-only transactions. Nomba is an expensive network to expand, requiring a huge investment upfront to build, as ‘agents’ need to be signed up and trained with the company not necessarily breaking even on them. Who are these ‘agents’ that the company revolves around? Yinka Adewale, the company CEO, has gotten the idea of agent banking from a customer who used the service in its early chatbot days to set up a shop where others could come in to send money. Initially thinking the high number of transactions from one account was due to fraud, the parties set up a meeting where this business model came to light. Seeing that people would rather come to this kind of shop rather than the bank which was standing literally opposite to it on the street, it became obvious to Adewale that what sells is quick and easily accessible money transfers through agents, which ultimately became the idea that allowed the company to turn profitable. 


The top-line is fuelled by charging a fee on every transaction, meaning the revenue calculation is a simple multiplication of the number of transactions processed times the average transaction fee. As mentioned above, the way those transactions come about is actually no longer purely AI-based, although the company was initially started as a dot ai startup. Hence, the service now is “a platform where merchants could use human interaction to process transactions, instead of artificial intelligence”. However, AI still plays a significant part in its operations. This is because it can use customer data to build a tailored solution for each business customer - businesses are now regular customers as the company has segmented its offering for small, medium and large businesses. Custom solutions mean having all business functions in one product: something like restaurants getting the capability of reviewing inventory, menu and processing payments. In fact, expanding beyond payment solutions towards invoicing and order management is now one of Nomba’s lodestars.

Unlike DataProphet, Nomba is not (yet) aimed at an international audience. Instead, it seeks to redress a local problem (of money processing) with a digital solution. It is a different use of AI in a noticeably varied context and the contrast here is a sound display of not saying that we ought to treat like alike: African AI startups may diverge from each other as much as Nomba and DataProphet do.








Stockshop.co.za is another South African AI startup. It is “an education, information and product platform” serving as “a free personal guide to the stock market where [customers] will be supported to learn, trade and invest”. The CEO, Annabel Dallamore, decided to start the company after working at a stock exchange and noticed that most people perceived the stock market as too complicated and opaque, resulting in the impression that it is hard to access. Thus, Stockshop was started with the intention to increase retail participation in the stock exchange, to make it “accessible for everyone”. It provides users with education tools (news, research, etc.) as well as services (share trading platforms, unit trusts, asset managers), and trading tools. There is also an online learning platform, the Stock Shop Academy, and Stock Shop Daily News, a daily newsletter. In regards to the financials, money is made by advertising  (lead generation) and taking a cut from trading fees. 


Besides the stock market offering, the company has branched off to offer an Africa-exclusive micro-insurance service, a conversational UI match-making financial brokers with client leads, and an algorithm connecting financial behaviour with emotion.


Stockshop again seeks to remedy a local problem, although it has seemingly evolved to tackle it at a much greater scale now and so reach out to a greater audience. It reminds us of how AI can be used in multiple ways even in young businesses, so the label AI startup does not necessarily connote a singular AI offering.



We have seen how three African startups are utilising AI in their main operations and/or offering an AI solution as their primary product. Notably, the startups mentioned are from South Africa and Nigeria, which is consistent with the concentration seen in the African startup space where the majority of startups are in Nigeria, South Africa, Kenya and Egypt. Besides this, the examples are indicative of how the AI startup scene is not at all unknown to the continent, which has seen these AI startups long before the current AI buzz: Stockshop was founded in 2013, DataProphet in 2014 and Nomba in 2016.

 

0 comments

Comments


bottom of page