“A scene straight from hell!” Is how one witness, a commuter, described it. Another report claimed it was the deadliest traffic collision in recent history. Diana (not her real name) was there. She saw it all to the split-second crash. Now floating on a terrible hangover with no previous fun, she wished she was close to her mother, at least aiding her to confront grief’s fogginess. Being the youngest of five, she was the most sensitive about family issues. She absorbed everything, especially her mother’s warm compassion and sound advice. But now she has to contend with the warmth of other suns because she failed to heed her last warning about travelling late to school. She wished she had done her back-to-school shopping at Blackberry Enterprises a day or two earlier to avoid getting caught up in the last-minute rush. Something was bound to go wrong, and sadly it was fatal!
She wished the lorry’s brakes hadn’t failed or lost control on the highway. Actually, she wished the lorry wasn’t there at all like the 52 innocent souls that disappeared into ether. The police cordon couldn’t hold off the hundreds who showed up at the Londiani junction. Despite the inclement weather, they flocked, everyone soaked with grief. Diana wished she had a chance to say goodbye to her loved ones before the unthinkable betided; to go back in time and change her story before it got relegated to the dark Internet’s underbelly. Knowing too well how people overpraise the dead and give justifications for their loss, and also realising that the impact of actions far outweighs any goodwill, she wished she could parrot out all her true feelings. Her star was undoubtedly on the rise; promising to shine all over the world with her artistic talent. But now the world will never know what it had, what could have been because her aperture is shut for good. And the only promise still echoing after her departure is of the Government introducing new road safety measures.
My eyes are welled up by the time I finish reading about the Londiani junction accident, helplessly reviving the soul, or rather thoughts, of a victim who died too soon. I am seated in a meeting room at OBG Virtual Office, Jadala Place, awaiting a guest to join a Zoom video call. Someone who has been audacious enough to pursue his vision and resilient enough for luck to look his way. Incidentally, I came across Diana’s story while preparing for this interview:
ME: Who is Arthur Mulwa?
HIM: Arthur Mulwa is an entrepreneur and problem-solver. There is no problem that I run into that I don’t feel like solving. My mind just loves patterns and puzzles. But most importantly, I am a husband and a father.
Arthur Mulwa — Co-founder & CEO of AI Care
Tell me more about your company, AI Care. What was the inspiration behind it?
That is a long, winding story but I’ll get to the tail-end of it. After spending a couple of years in the US and UK, working in the commodities trading industry, I decided to move back to Kenya hoping to land a gig in the capital markets space. During that transition, I met a friend who was working in insurance. He shared some of the frequent challenges he encounters in risk management and I realised that most of the algorithms or solutions existent in the capital markets arena are also applicable in the insurance industry. Prior to that meeting, I had already gotten a flavour for insurance; seeing it as a space I could make a positive impact in using my skill set.
I met my AI Care co-founders at a venture capitalists’ program run by Antler. Before our business in the insurance sector coalesced into something concrete, we had explored issues in several industries including healthcare. However, we lacked the professional prerequisite to tackle some of the challenges appropriately. As a team, we felt more confident in motor insurance because of our collective experience and passion.
Describe how you utilise artificial intelligence (AI) to underwrite motor insurance.
When we did a deep dive into the motor insurance industry, the “monster” we found was that 85% of underwriters lose money in the business. On one exigent arm, customers were also complaining about sky-high premiums. In short, you have a dissatisfied service provider and a dissatisfied consumer. That piqued our interest. It is a problem that can’t be easily solved without real-time accurate data. And that is why the foundation of AI Care is data-driven. We use high-performance computing (HPC) to process large amounts of data and analyse complex risk factors.
Our first product was mileage-based motor insurance: low-mileage drivers get to pay less premiums for having low risks on the road. Automation reduces discrimination and bias, so before launching the product, we analysed a large set of data on driving habits, especially on Kenyan roads. On our marketing side, we also analyse data to understand the kind of campaigns that we need to push to customers; how they respond to new and unique products or what they are interested in.
On road safety, what is AI Care’s main contribution?
As soon as we took off as a company, we quickly realised the significance of utilising our products to tilt the scales and change road safety outcomes. Not only would the policyholder get a better road experience, but it would also be more profitable for us. Studies show that over 90% of global traffic accidents occur in developing countries such as Kenya. And more than 270,000 people die in vehicle collisions in Africa each year. That said, we have studied aspects of driver behaviour that would make one more susceptible to accidents. For example, you incentivise good road behaviour when you price drivers based on their habits. Additionally, we have been talking with Government parastatals like The National Transport and Safety Authority (NTSA) in a bid to share relevant data to improve road safety countrywide.
How long has AI Care been operational?
We officially launched in June 2020 but we unveiled our first insurance product about two months ago. We also launched a product for asset financiers at the beginning of 2022. No grey hairs yet but we have gathered enough data and experience to authoritatively say that we understand very well what happens on Kenyan roads.
Is your company regulated?
We are regulated on different forums. First, as a telematics service provider. Telematics is basically how we are able to collect data on driving habits using devices installed in cars, and then analyse that data to get actionable insights. From that perspective, any telematics service provider is regulated by the Communications Authority of Kenya (CA). Secondly, any insurance product launched in the market needs approval from the Insurance Regulatory Authority (IRA). We are compliant in that regard and we have been working closely with not just the insurance regulator, but insurers as well. Finally, as a company that handles massive customer data in Kenya, we have to comply with the Data Protection Act. So, we are also regulated by the Office of the Data Protection Commissioner (ODPC).
Regulations are not put in place to hamstring; they are enforced to protect consumers. We ascend by tugging one another’s bootstraps, therefore all these regulations counterpoise to ensure that everything is above board; data is utilised appropriately and that customers know what their data is being used for.
Challenges you have faced as a startup in the Kenyan market?
We can talk about that all day! [Laughs] But I’ll just highlight the main ones. The speed at which big corporates and startups move are very different. You can see a big gap in AI Care’s timeline. We started working on these solutions in 2020 and we just launched our first insurance product the other day. I think the biggest challenge has been trying to move at the same pace as big corporates without dying in the process. We are externally funded but there isn’t an infinite supply of resources to pay for research and development (R&D) that we rely on to conceptualise and commercialise our products. To survive, we have had to make tough decisions. On the brighter side, being in “survival mode” has pushed us to be more innovative; optimise the little that we have.
Lack of funding, especially in this part of the world, is also a key factor. The kind of infrastructure required by startups to launch similar products isn’t available. You set out to solve insurance problems but end up solving telecommunications, fintech, and operations problems, among others. In more developed countries, these solutions already exist, so you don’t have to start from ground zero. And if you are trying to solve complex integrated problems, then the cost of the project or end product goes down or becomes more affordable.
How do you reach your customers? And who are your main clients?
Our solutions are B2B2C. To reach our end customers, we need a business in between to sell the product. That is basically the insurers. Our first objective, therefore, was to rope in the insurance companies. Takaful Insurance of Africa (TIA) was the first insurer we signed up with. Talks are still ongoing with Britam Insurance and a Ghanaian insurer. Hopefully, we shall launch a product with them in the coming months. Concerning the end customers, we have seen a lot of excitement. Motor insurance is renewed annually, so we expect to see a much larger intake of our products as we progress.
Do you have a product available on Google Play Store or Apple Store?
We have a website where you can access our first product: TIA SWITCH. Once you log in, you will be instructed on how to download the App on your Android device or iOS. The techno-financial revolution has made all manner of things possible and so everything is 100% digital, including payments.
Are your products tailored for specific markets?
One of the biggest problems in insurance is that products have been monetised to the point of one-size-fits-all. However, given the technology we have at hand right now, we want to ensure that we are solving specific problems for specific users. For example, low-mileage drivers already know that they are being overcharged. Of course, that is a big issue, especially during times when economic backdrops erode people’s appetite for savings or deductibles. We understand that every customer is unique and through data-driven solutions, we can add value to specific customer segments in the market.
Challenges and opportunities that autonomous vehicles pose to the motor insurance industry?
What was once reserved for the realm of science fiction is now very real. However, autonomous vehicles still have a way to go as they are not entirely effective in developed countries where roads are relatively good and road signage is perfect. Eventually, when both mechanical and infrastructural problems are solved, autonomous vehicles will enter the global market on a large scale. When that happens, companies such as AI Care will stand an excellent chance to provide valuable data to manufacturers on how to tune their models appropriately before commercialising.
At that stage, do you think the liability of road accidents will shift from drivers to manufacturers?
[Eyes narrow in deep thought] That is an area of debate across the globe. Various factors come into play when you look at accidents such as the environment, reckless driving, poor infrastructure, inaccurate algorithms, and more. So due to the complexity of the issue, legislative bodies will have to enter the fray and different countries will take different approaches. Another bone to chew will be how much autonomy we give to the machines because technology is not above errors. It is an integral question that can’t be answered in one go. It’s like a Jungian mandala. But I believe with the right amount of data we can come up with sensible conventions.
As a co-founder and CEO of an arguably game-changing tech company, what do you foresee in the Kenyan market when you telescope its adaptation of AI?
We have a lot of capable people in the country. By “capable,” I mean people who can function as solution creators and not just pieces of creation. So, I expect to see more data-driven, AI-driven solutions coming into the market. The advancement of AI is inexorable and thanks to the ChatGPT effect, many companies are exploring options on how to be more effective and efficient in their operations. However, AI is just part of the solution, not the full spectrum. Therefore, founders in Kenya still need to sit down and identify specific problems that need solving as opposed to the wholesale import of solutions. A transition is not a final destination, all in all, I think productivity will go up and the people who will benefit the most are the ones who quickly adapt and maximise the use of AI tools.
Leave a Reply