We are at the early phase of an AI (artificial intelligence) race with amalgam of platforms available for developers to build upon. We have Amazon Alexa, Google Assistant, Microsoft Cortana, Samsung Bixby, Apple Siri, and more. For African startups, the choices will not be easy because besides the technology, we need to consider primitives and datasets which can support an African business. In this piece, I examine some things to take into considerations as founders make decisions on the AI platform to adopt. You want to build a business on a platform your customers will like to use your products.
Around 2010, at the peak of Blackberry, I wrote that Blackberry would fade and Android would rise in Africa. I took many things into considerations when I made that call. I had lived in the U.S. with constant visits to Nigeria, and I saw some enablers and anchors which were in Android but lacking in Blackberry. It was evident that the adoption of the open framework of Android in a largely emerging region would happen.
Today, I want to provide the same direction for startups looking for a roadmap into the future of AI. Google has declared itself an AI-first company, and it’s doing everything possible to awaken the world with its computational capabilities. Amazon from nowhere changed the basis of competition with Alexa, a voice AI system which is evolving as an operating system for voice. With Alexa, new possibilities were unveiled and quickly incumbents like Microsoft got themselves in a catch-up mode. But this warfare is far from over: we are just starting in this sector since AI is still at the level of infancy.
So, as an African startup, you have a decision to make, since no one knows who the clear winner will be. We do not have the same level of clarity, at the moment, on AI as we have on Search (Google is the winner), social connection (Facebook is the winner), micro-blogging (Twitter is the winner), OTT (WhatsApp and WeChat are the winners), VOIP (Skype is the winner), and so on. Yet, AI will not have the same level of shape as most other sectors because AI itself is structured to make these platforms and ecosystems smarter. So, AI can improve Skype, WeChat, Facebook and more. We may not see a product called AI because there is nothing like that. That angle will help us to understand how to approach the adoption of a platform of choice.
But in this piece, I will narrow the AI into voice-activated AI systems. That classification is amorphous but I will go with it. Voice is going to be a huge factor in Africa since we like to talk and voice does not discriminate on literacy level. If you can build the solution in the native language of the people, you have a product that can be used in insurance, banking and other sectors. I have not included Facebook in this piece because it does not have an enterprise facing voice activated AI system. Sure, Facebook does AI but that is largely for its ecosystem, not for startups to build at scale through integration.
Here are factors to consider:
- Technology: The most important factor is that the technology works. If the technology does not work, there is no value even considering a specific platform. In this space, I give Amazon a lot of credit for its pioneering vision on voice assisted AI systems through Alexa. Yet, there is nothing Amazon does today that Google cannot do. Microsoft is also innovating. I will consider Amazon, Apple, Microsoft and Google as strong enough that the technology capabilities are largely even.
- Popularity/Network Effect: The construct of network effect is very important. You want to build on popular ecosystems, not just in U.S. but also in Africa. Google has an edge there because if Google Assistant on-boards on Android, you will have many clients since Android is popular in Africa. For Amazon, you will need to have a new hardware which increases the adoption cost. Apple is strong but not many Africans use Apple. For this, I give it to Google. Microsoft has Windows where Cortana runs but unfortunately, laptop and desktops are largely not as ubiquitous as smartphones.
- African Data: Among all the companies, Google is well positioned on data that involves Africa. YouTube had clearly documented African culture with our accents and lifestyles. In other words, Google Assistant can understand Africa more than any other platform, if it chooses to do so. It can understand how we talk, greet and stand in reverence to elders by reviewing videos we have uploaded on YouTube. No other platform comes close. So Google wins here.
- African Business Presence/Developer Engagement: Google and Microsoft have invested thousands of hours to grow and nurture ecosystems in Africa. These companies have organized competitions and sponsored conferences. Apple and Amazon are largely alien in building relationships with African developers and startups. Samsung is also local and working but Microsoft and Google are far ahead.
- Others: Google and Microsoft are strong on other factors. In most African major cities, you can pinpoint their offices. You cannot say the same for Apple and Amazon. Though Amazon is in the continent selling its Amazon Web Services, the focus is acceleration of consumerism. In South Africa, it operates a lackluster ecommerce operation where it has refused to invest more capital. So, Amazon seems focused on India as a place to win, neglecting Africa. Microsoft and Google do well on other factors. But Google has a clear edge because of YouTube which can help it teach AI systems how African humans hug, cook and drink through massive datasets of videos it has accumulated over the years. We have made those videos and Google can make sense of them if it finds market opportunities in the continent.
Google, which owns YouTube, announced on Oct. 19 a new dataset of film clips, designed to teach machines how humans move in the world. Called AVA, or “atomic visual actions,” the videos aren’t anything special to human eyes—they’re three second clips of people drinking water and cooking curated from YouTube. But each clip is bundled with a file that outlines the person that a machine learning algorithm should watch, as well as a description of their pose, and whether they’re interacting with another human or object. It’s the digital version of pointing at a dog with a child and coaching them by saying, “dog.”
I must note that Africa does not have any of the core technologies like Alexa, Siri and Cortana to build AI solutions. It may not really be absolutely necessary for us to pursue such goals at the moment. We do not need to have Windows before we can create products for PCs. Though I do hope we can get there, the fact is that we do not have the financial and technical capabilities at the moment to pursue such expensive endeavors. Note that I did not say that it is impossible: I am saying that if we do it, we will not win because Silicon Valley will out-compete us. Our opportunity is to find a space to differentiate on these platforms locally. Yes, we can solve local problems using these platforms. If that is our focus, I do think that local data will be catalytic to success. In that space, Google wins because it has more data about Africans and Africa than any other company. And for that, I will suggest that you consider building on top of Google platform because it will win in Africa even if Amazon Alexa wins in U.S. Yes, it will mirror the smartphone operating system platform where Apple is winning in U.S. (across many metrics and indicators) but Google is doing better in Africa.