Identity verification is the new talk of the town. After Elon Musk made it a paid service exclusively for Twitter Blue users, Mark Zuckerberg has followed suit by announcing this week that identity verification will become a paid service across Meta platforms including Facebook and Instagram.
Smile Identity, a pan-African identity verification startup recently raised $20 million Series B to ramp up their products across the continent. The startup has been building their product since 2017 and are ready to take to new limits with the new funding.
TechCabal caught up with Mark Straub, CEO of Smile Identity, for a discussion over a wide range of topics including raising funding during a downturn, the importance of identity verification in accelerating the digitalisation drive in Africa, and much more.
TechCabal: Please tell us more about Smile Identity and the problem that you trying to solve
Mark Straub: The thing about doing business over the internet with consumers is that if you’re a business, at some point, if you’re touching money, you need to know who you’re dealing with. And you need to be able to confirm that you’re dealing with a person who is who they say they are. And that has been a difficult thing to do for a long time across many markets in the world, not just in Africa.
Facebook made this announcement this week that they’re going to start charging for verified profiles, but they’ve spent probably the last half decade trying to confirm the identity of accounts inside their system and make sure that they’re not opening accounts for bots or other fake or fraudulent profiles. And so the idea of the problem of solving identity on the internet is actually a really global and universal one.
It just so happens that it’s very acute in certain African markets, because of the limited availability of datasets or infrastructure. And that is now changing with the rise of new national ID systems, and increasing usage and adoption of a variety of different kinds of forms of ID.
And so our ambition at Smile Identity is to make it easy for people to prove their true identity on the internet. By just being who they are, whether that’s taking a selfie,typing in their phone number, or ID number, or taking a photo of their ID card, we want to make it really easy for people to prove they are who they are, so they can get access to the digital lifestyle that they want. And we combine a bunch of different tools to do that. At the core of it is integrations to local data sources, essentially sources of truth. We combine that with biometrics that we’ve tailored to work well on devices and network connections across Africa. And then we also combine that with our document verification, so that we can verify the legitimacy of a wide variety of documents across Africa and beyond.
So we bring all those things together in a set of APIs and SDKs, and a developer dashboard and Docs. And we try to make it easy for developers at financial services companies, at ride-sharing services, or even at banks to confirm the true identity of their end users. And do so while onboarding users at scale cost, effectively.
TC: What would you say are the most pressing challenges that you have incurred trying to solve this problem?
MS: There are a lot of different challenges. I think they range from what I’d call sort of the infrastructure side. So you have some data sources that go down periodically, even national ID systems that go down maybe as frequently as once a day, or once or multiple times a month. And so we have to find ways to mitigate that. Sometimes it’s by building automatic retries. Sometimes it’s by having redundancy to enable multiple connections to different data sources.
We also have to deal with the device landscape so that the wide variety of handsets and operating systems and network connections can work properly. So the infrastructure side is one real challenge, and I think everybody has to deal with that. But when you’re building tools that include capturing images, uploading files, hitting up different APIs, and then bringing that all together in an asynchronous fashion, that can be really challenging.
So that’s one set of problems that we’ve tackled. And we’ve tackled it by hiring mobile engineers, and by building custom software for every combination of operating systems and handset. And then the other challenge I think is understanding local expertise and compliance across multiple countries. And so building out that footprint has taken us time. But we now have teams in Cape Town, Nairobi, and Lagos. We have individuals in a handful of other markets like Rwanda, Uganda, and Ghana, and we’re now beginning to spend time with people that we are probably going to hire in Francophone markets.
So understanding the local context and making sure that everything we’re doing from a data standpoint, is consistent with local data protection laws has been a bunch of work. Getting that right requires a lot of coordination across lawyers, compliance rules, software, licencing, and in some cases, incorporation in local markets, or having to have directors or boards of directors in those markets. But we’ve really kind of embraced that and said we want to meet the highest standards for compliance across the continent.
And so we’ve really invested time and energy and human resources to make that a part of our company. And that’s a lot of work to get it right but it’s critical if you’re going to build a company that’s all about building trust in the continent.
TC: Smile Identity recently raised a $20 million Series B. How difficult was raising such a significant amount of money in a VC downturn?
MS: It was hard. Fundraising is always hard because you’re asking people to part with large sums of money. I think the part of the story that allowed us to be successful was the amount of growth we experienced in the last 12 months, which was significant. We grew the business into three-fold. The other thing is that the problem we’re solving is a very long-term one.
So while the immediate environment this month, or this quarter, or even this year may be challenging, the long-term trend towards the digital economy is not going to stop. And people are not going to go back to using paper or going to bank branches to open up bank accounts. They’re going to continue to do that via digital channels. I think investors agree with us on that hence their backing.
TC: What will be the main use of the funding that Smile Identity has just raised?
MS: We’re expanding the product engineering team, both in Africa and globally, and investing more in our biometrics, and customising and localising our document verification solutions. We’re finding that there’s a lot of nuance in handling documents in certain African markets. So for example, in Cameroon, the document has a green number in one shade on a green background. Capturing that correctly is really important. In Senegal, sometimes the IDs have a certain number on the front, sometimes it’s on the back. In other markets, the IDs have QR codes that you can use to check whether the textual information on the document has been altered or not. So handling all these little nuances in each of these markets is really important.
That’s the localization of the product that we’re doing. We’ve also begun investing more heavily in things like deduplication. So that’s a computer vision solution where we can check to see whether the same user has signed up for your services multiple times. We call that solution ‘Smile Secure’. And that’s a big area of technology and AI investment for us.
The second area of investment for us, which is related to what I was talking about regarding international expansion; we’re building new teams in Francophone Africa and beginning to explore North Africa as well. And in those markets, of course, we have to handle French and Arabic. As you get into optical character recognition, and local language support, that becomes a fair bit of work. And so we’re going to be doing that work.
The third area, which I think is the part that maybe is most interesting to your audience, but I think most aligns us with the whole ecosystem is deepening our investment in identity infrastructure. And what I mean by that is, we’re now starting to see more data protection laws with specific consent rules. So what we are doing is building those consent rules into our SDKs and consent layers that users will see.
This means that if you are a bank or a lender, and you want to start using our services in a given market, like Uganda or Nigeria, instead of having to figure out how to build all that consumer consent into your product, you’ll be able to integrate our software, and know that you’re compliant from day one. And you’ll have consent receipts that essentially allow you to prove it. Our software will also power that without you having to go figure out that user experience. So we will basically allow enforcing local data protection laws with pre-built consent layers and consent software, that fulfil local requirements. That’s a really important area of local investment.
The last piece about local investment is that we’re continuing to share anonymised metrics and insights about fraud trends across Africa. We did this with our KYC report which we announced a couple of weeks back. Those KYC reports, and the insights that we’re generating, will continue to develop and expand and share. And as we get more and more data, we obviously see more and more types of attacks. And we also see more ways of preventing attacks and protecting consumers. So those are the three areas we’ll be investing in.
TC: How does Smile Identity try to ensure that its algorithms are not biassed and end up discriminating against a particular demographic of users?
MS: The first thing is that we’ve been training our algorithms almost exclusively on faces across Africa. So while we have a few users globally, we have a heavy training direction towards African faces. And so , just by the nature of the modelling that we’re doing, we’re not modelling on Eurocentric datasets. So that just makes a big difference. But the tricky thing is you want to make sure that your system should not be biased too much in either direction.
So we wrote a paper [pdf] about this. We basically looked at our dataset, we looked at a couple of other publicly known datasets and we trained and compared our dataset and our algorithms. What we found was that we were able to outperform some publicly available algorithms on African faces.
We found that if you were, let’s say, verifying 100 users, one service would be able to get you 98% correct results, we might be able to get you 99%. Or if another service was able to get you 99%, we might be able to get you 99.8%, or 99.9%.
TC: What role do you think products like the one being built by Smile identity will play in accelerating the digitalisation drive in Africa?
MS: We see ourselves as an on-ramp, and an accelerant towards the digital economy that we know is coming across Africa. I think there’s a great quote by technologists who said, “the future is already here, it’s just not evenly distributed”. I think you can watch Nigerians on Twitter and see that there are many developers and founders and engineers who are very much living in the future.
We think that we can accelerate that adoption of the digital economy and we can do so by making identity verification work, whether you’re in Cote d’Ivoire, or Senegal, or whether you’re in Ghana, Botswana, and not just for people. We think we can basically help distribute the future and accelerate the future across the continent. And we were doing that by making it easier and easier to use our technology whether you’re dealing with a mobile phone, a website, an app, a tablet, or even an analyst who might be sitting in a back office and just needs a tool.
TC: What would you say sets Smile identity apart from the competition in the identity verification space?
MS: We’ve been doing this now for half a decade. And we’ve done it at scale and we’ve done it across almost every industry that I can think about, from ride-sharing, to banking to even social services during COVID. And we’ve done it in over a dozen countries where we’ve actually done some form of verification. We’re also compliant and meet the local requirements and the local laws and all those markets.
There are a few local companies who have data access in one or two countries, maybe three countries in Africa, but they don’t really have a solution for handling the whole variety of documents that we see across the continent. What we pull together is a combination of biometrics that specifically works well on the continent integrations, methods that work well across the continent, and local data sources wherever they’re available across the continent.
But we’ve also combined that now with document verification solutions that work across the continent. So if you’re looking for a pan-African solution, that both work well with high pass rates, and also is compliant with local data protection laws, there’s not really another company out there that brings all that together.
TC: What trends do you see being prevalent in identity verification over the next two years?
MS: Well, it’s interesting that we’ve just seen this announcement from Facebook about the idea that they’re going to create a meta-verified profile so people can pay to have their identities verified. Of course, this is picking up on the idea that Elon Musk introduced at Twitter about verification on Twitter Blue. There’s a real cost associated with verifying the identity of people on the internet and so far, that cost has mostly been borne by applications rather than by consumers directly.
It’s been interesting to see these two giant social networks now begin to try to offer that verification solution as kind of a paid service. So I’m intrigued to see whether that trend carries over in Africa and whether people who want to have, let’s say, access to higher loan limits or access to international cards or access to higher transaction values, will be required by the application providers themselves to pay for verification.
I also think we may start to see KYC becoming a thing that you don’t do once, but you do it kind of periodically as the amount of risk or as the amount of services in that account increase. And maybe that will begin to drive companies to offer consumers the chance to pay in order to accelerate that movement through the curve.
So instead of having to wait six months or a year to get verified, and then to get access to some more sophisticated service, maybe consumers will sort of be given opportunities to step up their KYC upfront, but in exchange for some additional cost.
*Interview has been condensed for clarity