BRIGHT SIMONS is the president of mPedigree, a social enterprise working on three continents in partnership with governments, Fortune 500 companies and grassroots organisations to promote innovative technologies, from sensors to software, in health & food security, transparency and governance. In this interview with KEMI AJUMOBI, he shares on the importance of data integration and how Nigeria and Africa as a whole cannot afford to lag behind. Excerpts
How important is data integration in African businesses?
There is an erroneous impression that data-driven decision-making in Africa is a luxury or a distraction due to the constant uncertainty, low availability of trained talent, and the “primacy of the mundane”, i.e. the sheer number of “basic blocks” a business has to clear on an hourly basis just to stay in business.
This is not a sound way to look at things. I see many instances where data-driven decision-making is not only directly relevant to the bottom-line but also critical to survival. Take inventory forecasting for instance. How often have people not gone into a clothing retail outlet to find that all the clothing are of one size, and not the right size for them alas! That’s a classic case of businesses not being able to use historical data to plan demand properly. Even the long service wait times most people complain about is often due to weak demand forecasting.
So, whilst I argue in my CGD paper (A Farewell to Disruption in a Post-Platform World) that many businesses in the West are increasingly struggling to harness data due to fragmentation (data being splintered across many control points), and most of the low hanging fruits having already been plucked, in Africa, there are still many elementary opportunities that we are not really taking advantage of. Especially considering that data-driven decision-making can pay for itself in the short-term.
African countries are fast embracing the importance of data and making regulations to that effect. What is it about Artificial Intelligence (AI) that they need to embrace speedily and what do they stand to gain?
I see some great developments in the banking sector in Morocco. Due to huge pressures on profit margins, the South African mining sector is starting to seriously embrace data science. Renewable energy virtual grids in Kenya (such as “pay as you go solar”) have always believed that data is salvation. So, the usual bright spots. Unfortunately, we don’t have a lot of national-scale success stories on the continent. AI, at this point of its development, is largely statistical and highly reliant on the efficiency of data flows and ease of consolidation. The only sector in Africa that is attaining the degree of interoperability necessary for AI to start having a serious impact is the financial services sector. As always, there are low hanging fruits for places like Africa to emphasise. Credit scoring and credit risk management are obvious choices though once those early gains have been creamed the obstacles to generate more marginal value are formidable. At the current state of business ecosystem design around the world, diminishing returns set in faster than one would assume when it comes to AI because the protocols to allow data from truly diverse sources to converge within AI applications are still being built. A lot of AI is thus incomplete AI, and incomplete AI is problematic. Remember the old saying: a little knowledge is a dangerous thing.
Using Nigeria as an example, how beneficial can AI be to a highly populated country with enormous advantages and opportunities, using their available resources to buttress your claim?
Government influence on industry design in a place like Nigeria is far stronger than in a place like the US or most European states. Government seriously has capacity to influence interoperability and cross-portability standards, protocols, and frameworks across major vanguard industries like financial services and modern retail. The ability to get to more complete versions of AI in areas like credit risk assessment and predictive sales planning is much more pronounced. Even better, government agencies can be brought into these “honeycombs”, as I call them. Nigeria can weave more complex tapestries of public-private technology platforms than is possible in the West because the private sector genuinely looks to government for forward-looking intellectual leadership. This is far from the case in the West.
How important is it for governments to create enabling environments that can inspire local businesses to be efficient and productively inventive?
Right now, “enabling environment” means very specific things. No more abstract macroeconomic jargon and broad targets for inflation, exchange rate depreciation and interest rates. I am not saying that these are not important. They are but they have become minimal standards internationally. The focus is back again on “infrastructural scene-setting” and “institutional quality”, but the notions of what we mean by “institutions” and “infrastructure” have become more eclectic. Folks want government to create innovative infrastructure coupled with innovative institutions. Mostly prototypes. And the primary purpose is to enable individuals and companies to learn and unlearn fast. So, whereas in the past, it was all about how to build institutions and infrastructure to enable stability, citizens now want to get to a point where they can take stability for granted so that they can judge governments on the real deal: how creatively they can layer on top of the stability new mechanisms for learning fast. For example, right now, digital capabilities are moving too fast for traditional thinking about educational policy, even by competent technocrats, can catch up. Prototypical systems are required to reinvent credentialing and rethink curriculum design and delivery. This is partly because even more of the talent and skill tooling shall happen within work settings than within school settings. A government that cannot test new ways of creating portability across academic and out-of-school certification isn’t worth anyone’s time.
What is your take on disruption and disruptive innovation? What is your attitude to both and why? How does your approach positively benefit Africans?
I can sum up the current chief megatrend in technology innovation as follows: everything is weaving together, but the joints are not seamless. Data sources are more fragmented than ever before and the data itself is more heterogenous than ever because a very broad variety of entities are now generating data that impact on the accuracy and utility of data held and used by other entities. The broad span of modern technology systems can be simplified along three dimensions: data, algorithms and integrations. Algorithms are generally more persistent (they are useful for longer stretches of time) whilst data has become increasingly very transient because data utility degrades very rapidly. But algorithms are also getting super specialised as we learn to do more and more previously manual stuff with more technology. This has made “integrations” the most important factor in techno-business today because it is how more and more algorithms can be efficiently connected to fresher and more varied data streams rather than the algorithms and data on their own that dictates value creation. This process is reducing the power of single companies to make radical breakthroughs and to disrupt the status quo as at any one time an innovator needs more consensus than ever before to change the system. My thesis in the CGD paper is that, this has made “disruption” far less attractive to top innovators than “smart alignment”. At the same time, it has blocked many of the “disruption paths” available to underdogs to rise to the top. Because the African innovation sector is made up primarily of underdogs, the picture has actually worsened for Africa’s ability to utilise innovation to break out of its global marginalisation unless we create special infrastructure that reduces the burden on individual innovators in connecting powerful new algorithms to rich, emerging, data sources in order to build critical solutions in sectors such as health, education, energy, and tourism which are still relatively open for the taking because they have yet to be colonised by the platform megatrend dominated by the West.
What is your opinion on the convergence trend. Would you say the economics of different industries especially in Africa are not necessarily converging?
“Convergence” did not really happen in the way that was anticipated. The initial hope was for technologies to become so multifunctional that they spread uniformly across all industries and reshape their fundamental economies to fuse their impact. What we have seen is, indeed, greater connections among industries but without their fundamental economics really converging by as much. There are still laggard industries where digitisation is failing to seriously deepen productivity. Education, health, agriculture, and even energy and transport. Attempts by pioneer companies like Uber to force convergence have so far seen generally lacklustre results and undermined their profitability and thus capacity to enforce hegemony. This is a worldwide trend. And it is a trend that actually gives Africa some breathing space to make a mark in some of the laggard industries. As I argue in my CGD paper, instead of convergence we got “hyper-integration”. A more viscous/friction-fraught type of convergence. It has saved African industries from becoming mere outposts of Western digital hegemony. But only for a while, and only if African leaders are smart enough not to miss the latest boat.
Share on integration inflation and its influence on Africa as a whole
To recap: “integration inflation” refers to the situation where technology-enabled products and services need more and more connectivity to other parties’ products and services in order to deliver value. The result is an increase in cost, time and friction needed to generate innovative products and services. In Africa, it has made it very hard to extend the fintech boom to far more critical sectors such as pensions, insurance, and mortgages. It has more or less snuffed an early spark of e-health and edtech glory. If “pre-fabricated” integrations infrastructure is not constructed as public resources, African innovators will find it very hard to take advantage of the current window of opportunity to transform laggard industries like health, education and agriculture where Western digital forces are still struggling to completely dominate.
How does Technology, Entrepreneurship and Innovation (TEI) power economic development? How can Nigeria benefit from this?
Like most African countries, Nigeria’s biggest hurdle now is how to improve the productivity of the teeming masses of underemployed and underutilised talent. Technology, Entrepreneurship and Innovation (TEI) shall play a huge role in uplifting the productive potential of these vast populations, the overwhelming majority of whom shall increasingly be under 35 of years as the century progresses. As Arthur Lewis saw as far back as the 1950s, the potential boost in the productivity of underemployed labour is by far the biggest potential boon this continent can harness to shift its economic status within a generation.
Rate Fin-Tech in Africa and its development. What can be done to improve the growth, what are the rewards?
Fintech is one of the few areas globally where digitisation has succeeded in seriously converging multiple legacy industries into an emerging mega-industry globally. Africa has quite successfully jumped onto this global trend, mostly because more than 60% of effective TEI funding over the last couple of years has gone into fintech. Fintech’s experience in Africa show clearly how digital transformation can accelerate when the right trends align. Unfortunately, there is a serious misconception that fintech by itself can seep into multiple sectors and transform them. Other industries need their own transformation before they can interlock with fintech trends. That is why, hopes that digital payments by themselves would make a huge impact on the fortunes of e-commerce and e-logistics has simply not been realised. In many parts of Africa, many consumers still prefer cash to pay for ride-sharing services, for example. In essence, fintech has had very little effect on the growing digitalisation of distributed mass transit in Africa.
Awareness of a need to shift government technology transformation thinking from discrete G2C (government-to-consumer) and G2B (government-to-business) is growing. What’s your view?
It isn’t too surprising that when most African governments perceive digital government to mean the placement of electronic forms on the internet, to enable access to discrete government services like tax filings, passport and drivers’ licenses applications, and wage slip notifications, these are surface developments. They improve convenience for citizens, which is a good thing but hardly transformational of the essence of government itself. Transforming government will require improvements in accountability, agility in the delivery of government services, reductions in arbitrariness and abuse of discretion, and far improved capacities to use sensors to adjust government actions and spending in near real-time. In fact, the possibility of much more differential taxation based on highly personalised taxpayer profiles etc. are far from being well conceived, not to talk about even discussed. True government transformation via innovation is therefore, clearly, one of the prizes still dangling from beyond the reach of many countries, including most of the ones in Africa.
Why do you feel that development problems confronting the planet will in time only be addressable through such hyperlattice structures especially in developing countries like Nigeria amongst others?
I feel so because fresh, constantly changing data, from tremendously diverse sources need to be fed into super-specialised algorithms (computer programs) being developed in many segments of global society, whose problems are only now becoming transparent to the digital method. A new “world architecture” of cyber-enabled problem-solving is urgently required. Expecting standard public and private sector collaboration and partnership models to weave together the necessary inter-relationships across firms and government agencies, not to talk of NGOs and religious bodies, is to leave everything to chance. But in an increasingly compressed world, such institutional nonchalance is quite risky. There is a need to invent wholly new forms of institutions to deal with the fallout of all this data and algorithmic intensity. I have coined the word, “hyperlattices”, to reflect this new urgent imperative. Developing countries have an even more urgent need to take control of these hyperlattices that are emerging in multiple arenas such as travel & tourism (think the fusion of aviation, accommodation, and government clearance), trade, financial services and energy and infuse them with developmental aspirations. I call such positively mutated hyperlattices, “honeycombs”.
How does meshing high-digital and low-digital operational models to create new business models make it harder for platforms to externalise their risks?
When platforms like Facebook, which made their impact through enhancing digital communications and social networking discovery, venture into fraught, previously low-digital, terrains such as high-stakes national politics, their latitude to leave actors in the medium they have created to sort out the fallout outside their platforms becomes increasingly untenable. That is why the big tech platforms now employ tens of thousands of risk analysts and content moderators to directly deal with the effects of hate speech, fake news, cyber bullying, and assorted infractions on the civil order that in the past they left to the algorithmic orchestration of the digital social networking model itself to sort out. Note however that this seriously increases the barriers to entry for other players. In fact, one of Facebook’s frequent arguments for why WhatsApp couldn’t have thrived outside Facebook is the burden of internalising the full spectrum of risks generated by meshing.