Q&A: Onfido on the future of player verification

EGR Intel: Can you tell us a little bit about Onfido and the work it does?

Husayn Kassai (HK): Onfido is building the new identity standard for the internet. Our AI-based technology assesses whether a user’s government-issued ID is genuine or fraudulent, and then compares it against their facial biometrics. That’s how we give companies like Bitstamp, Small Screen Casinos and Trada Casinos the assurance they need to onboard customers remotely and securely.

EGR Intel: How does what you do differ to your competitors?

HK: At Onfido, we use a supervised, or ‘hybrid approach’ that combines machine learning with human experts. It’s the optimal approach for businesses looking to scale up their onboarding. A 100% automated solution delivers speed, but will also reject any users it can’t process – meaning you lose a lot of genuine customers. A fully manual solution gives you a better pass rate, but is slow, and not scalable. With our hybrid approach, any users that can’t be processed automatically are passed to our expert team for review – providing additional uplift rather than rejecting potentially genuine users. Moreover, the expert team trains the machine learning models at the same time to achieve continuous improvement.

We’re also document specialists, with a world-class team of leading document fraud experts. By adding a document check for any new users who don’t pass an initial identity check, some of our clients have been able to increase their pass rates to well above 90%. In a space as competitive as igaming, it can make all the difference.

EGR Intel: What is the technology behind this and how is this changing?

HK: We’re seeing more and more businesses talk about AI or machine learning, but for operators, it’s often not clear what that means, how it can drive value for them, or how much to believe. Our ultimate goal is to help our clients get as many genuine users on board as possible, as quickly as possible – and to make sure they keep bad actors out in the process.

At Onfido, we specialise in machine learning (specifically, deep learning). We use it in three ways: the first is OCR extraction, which automatically extracts the data from identity documents to pre-populate a sign-up form, for instance. This removes a lot of friction for users. The second is to look for signs of fraud in identity documents. This provides a greater level of assurance for businesses, as a lot of fraud won’t necessarily be visible to the human eye. The third is to read facial biometrics.  Our tech is able to match a person’s face to the face in their identity document, and to check that the person is actually present and alive – so fraudsters can’t cheat the system with a photo, for instance.

EGR Intel: Are there any inherent challenges with regards to verification in online gaming you think need addressing?

HK: Young male gamblers present a lot of challenges. Because of the close association between sports teams and betting, gambling is normalised from a young age – so operators see a lot of underage boys borrowing IDs or details from friends and family to get access to their platforms.

Problem gamblers are also tricky. When a problem gambler asks an operator to prevent them from continuing to gamble, it’s up to the operator to ensure that they don’t open another account elsewhere – which they can easily do using a fake name, address and date of birth. The same goes for bonus abusers, who’ll use multiple accounts to get free bets on each.

Using a document check can help solve these issues, as it gives operators more assurance that the person trying to join their platform really is who they claim to be. It prevents them using somebody else’s ID, and protects operators from the enormous fines that come with non-compliance. It also means that where users are genuine, you don’t lose them. Younger players might not always appear on credit bureaus as they don’t have a well enough established credit record – using a doc check means you can still get the legitimate ones verified and onboarded.

EGR Intel: Do you think onboarding will be one of the key customer battlegrounds in online gaming in the future?

HK: Definitely. If the new age verification laws are passed, things will get a lot tougher for operators. At the moment, it’s possible to onboard users first and verify their age later, when the user might already have spent a lot of money on the platform. With these new laws, they’ll have to verify age at the point of onboarding. Inevitably, that’s going to add friction – and we know that most gamblers will go elsewhere if  their user experience isn’t good enough. But fail to perform checks thoroughly, and operators face increasingly steep fines that in some cases, could put them out of business.

Finding a way to deliver security without compromising on speed and convenience will be the key problem for operators to solve in 2019.

EGR Intel: How are the new age verification rules going to impact onboarding for operators?

HK: Operators are going to have to rethink the way they onboard and verify their users – but actually, they can play this to their advantage. We’re seeing a move towards more positive identity verification in the market, and it’s important to recognise the value this can add to a business.

Age verification at the point of onboarding will add friction, but it’s also an opportunity to improve pass rate and protect the operator’s platform. It’s in operators’ interests to be thorough with the checks they perform, as non-compliance will cost them huge amounts both reputationally, and in terms of revenue. Getting it right now will drive value in the longer term, so the smart operators will embrace this change as a benefit, not a blocker.

EGR Intel: One of the key challenges in onboarding and retaining players is cutting out friction – but that doesn’t only happen at the point of onboarding. Where else can technology be deployed to help improve the user experience?

HK: Onboarding is important, but it’s not the only place you can use identity verification to your advantage.

Take chargebacks, for example. If you ask users to perform a document and face check at the point of deposit, they won’t be able to claim it wasn’t them that placed a bet. Account Takeover is another use case. Currently, operators will hold high-value payouts until they’ve performed a more through identity check on the user. That can take days, and is incredibly frustrating for your genuine players. A document and face check at the point of withdrawal is quicker, more secure and less resource internally. It keeps your users happy and your platform safe.

Husayn Kassai is CEO and co-founder of Onfido. Onfido helps businesses digitally onboard new users by verifying their government IDs and comparing them with their facial biometrics using machine learning. Founded in 2012, Onfido is now a team of 230, has received $60m in funding, and works with over 1,500 fintech, banking and sharing economy clients globally – helping them onboard more users while cutting costs and risk. Husayn has been named as one of Forbes’ ‘“30 Under 30”’, and as the World Economic Forum’s ‘Technology Pioneers’. Husayn sits on the Advisory Committee of the Oxford Seed Fund and the All Party Parliamentary Group on AI.