Skip to content Skip to footer

First world F-AI-L:

Why first world software solutions, specifically A.I. solutions, often fail in Africa

A thought piece by Dave Keating, CEO & Co-Founder of Refraime

Although Africa is a continent that is very prolific in producing leading edge innovations in the software engineering space, often including world firsts, in my opinion we far too often look to first world products to solve some of our continent’s problems.

We seem to always look outward for global technology trends, identifying products, services and providers that play in the latest and greatest arenas.

In my software career spanning nearly 40 years, I know of, or have been close to, a number of projects that selected so called “first world” products and providers to bring “new” functionality onto the African continent with less than optimal results.

Here is my take on a few things that can cause these technologies to fail in the broader African context;

  • In many instances, businesses assume that a new product or technology is appropriate in an African setting. They then bring the product or technology into Africa, only to find that their assumption has been incorrect, due to any number of issues that pose barriers to entry. Examples include legal implications in our jurisdiction as opposed to the product’s domicile jurisdiction, the maturity of the target market or even the actual need for, and uptake of, the product or service.
  • First world solutions inherently assume first world infrastructure. We are all currently experiencing the impact of loadshedding in South Africa and the effect it has on our computing infrastructure and networks. Often imported solutions require an inordinate amount of compute resource as well as permanent and robust online connectivity at generous bandwidth speeds. This is not realistic for large swathes of Africa and the consideration for offline functionality in our context often goes unnoticed. Obviously large infrastructure requirements dictate large costs which sometimes mean that the product will not be feasible from the outset.
  • Localisation can be an issue when importing technologies from other regions and, in the past, systems have been released into the market without any localisation having been implemented by the provider. This has caused the product to fail as well as, and more importantly, the underlying enabler to be discredited, resulting in rejection of useful technologies by the community at large. This has been particularly prevalent in African markets with facial recognition systems controlling enrolment and security access to systems and environments https://sitn.hms.harvard.edu/flash/2020/racial-discrimination-in-face-recognition-technology/ . The problems caused by lack of localisation have been even more pronounced in voice recognition systems. When voice became the next big thing in the late 1990’s and early 2000’s, automated IVR’s and voice attendants for everything from airport parking access to voice activated lifts https://www.youtube.com/watch?v=HbDnxzrbxn4 sprang up all over the place in South Africa but, when people could not get them to recognise South African spoken English (in the myriad of accents our population holds) the systems were quickly abandoned, disabled or overridden. These problems, although no indictment on the underlying technologies, caused the use of voice, which is probably the single most useful human factor that enables technology in the most ubiquitous way, to fall out of favour for a long period of time. This is a huge disservice to under-served African populations who could have leveraged their voices to access all kinds of services which could have been enabled via the microphones on their telephones. Considering that the continent is a long way from having every citizen own a smart device, this technology could have played an enormous role in uplifting and including people in the formal economy and government.
  • Lack of competent local support can be a contributing factor in preventing systems from gaining general acceptance and going on to be successful. International players often only establish a local sales and distribution presence in the relevant African market. They will not invest in staffing and training of local people to enable proper support and localisation of the technology, choosing instead, to do this from their home base or from an outsourced region such as the Asian sub-continent or Eastern Europe. This causes breakdowns in communication and delays in providing the type of support that business users require. In the industry we often hear of users having to wait for a resolution from someone in a time zone somewhere on the other side of the world. This, in my opinion, is arrogant behaviour by multi-national corporations who simply want to reap financial rewards from the continent without doing anything to uplift local communities and should not be condoned or supported.
  • Provider selection is often driven by a very conservative and risk-based approach for large corporates. When a limited number of players dominated things, an old saying in the industry used to be that, “you couldn’t get fired if you used the dominant one, regardless of whether the project failed or not.” This approach sometimes means an international or multi-national provider will be the de facto incumbent. This in turn dictates the product selection which may suffer from one or more of the problems already highlighted. It also means that local providers are overlooked and potential revenue and jobs are lost to our economy.

So, from all of this, my takeaway is that many of these pitfalls can be avoided through proper assessment of business needs and how to fulfil them.

This should include a formal assessment using performance acceptance criteria (documented before the assessment) for the applicability, cost, capabilities and limitations of any new technologies or platforms. This needs to be done together with an assessment of the provider of the technology to assess if they are fit for purpose.

Our problems and challenges are uniquely ours and Africa has loads of seriously good talent in the ICT space which is far too often overlooked by potential local customers. These people and solutions are ideally positioned to provide and/or produce truly African solutions for African problems. Keeping money on the continent and growing local talent whilst creating jobs is an imperative for our survival and growth as far as I am concerned and local really is lekker – you should try it sometime. 

Footnote

Not all technological advances are good for a lift-and-shift approach from their domicile territory into other territories. I have focused this thought piece on the First World to African transition scenario because, after all, this is where we live and in my opinion we see these failures more often