Abstract:

Digital Identity has well earned its place in the top 10 technology trends in the world. Amidst unprecedented times brought on by a global pandemic frictionless access to digital services has become more important than ever. Large corporations, governmental services, small businesses all alike suddenly got caught up in situations where well entrenched onboarding services even though supplemented well with digital elements involved could not cope up with a demand to go completely online. Reducing barriers to entry meant compromising on secure and validated access to services. This is where digital identity tech shone and became more relevant than ever. The goal was to integrate it in business processes maximising efficiency, avoid need of physical contact without forgoing validated and secure access for consumers to their services.

At Arrk, we have enabled one of our newest customers to unlock exceptionally well, the value in digital identity to create a compelling value proposition for their existing suite of products. Thus staying ahead of the curve before the pandemic would change equations majorly in the identity technology world.

The world of digital tech can throw many confusing solutions and products so much so that the focus of ensuring product market fit get shifted to soon poring over these varied applications readily available in the market. The key focus was to ensure we understand the business context of the problem well, develop an apt value proposition and choose the right technology fitment.

Context: The idea was to offer a proof of age product to consumers along with the incumbent product suite. The original process to be an issuer for such a product only included an online application and payment process but majority of the verification components of the process required the consumer to get their identity verified in person with a verifier. This meant that the framework even though a useful tool for retailers to combat underage fraud for access to their products had a significant barrier to entry.

Discovery: As part of the research for this proposition, product owners from Arrk worked closely with business stakeholders owing the product vertical and at equally with auditors from the compliance organisation that controlled the framework. Using iterative discovery techniques we uncovered existing pain points in the process with current stakeholders and made well accepted proposals around transforming the current process into a completely online process that would allow applicants to complete the process. The resultant product would go on to be also the first implementation of a 100% digital solution for the compliance organisation with our customer paving the way for the designed framework to stay relevant and viable in the future too.

Proposition development: The proposition development phase then focused most of its energy on ensuring proper product market fit. As-is process mapping of customer needs and pain points in their journey to access services from retailers and establishments was conducted with experience interviews with customers. This helped tailor the onboarding journey to cover users who would opt for this special product. The challenge was to ensure no messaging confusions for users who had no interest in the new proposition and not alter the common onboarding journey significantly. Two main components of the use case of digital identity verification for any business are:

1. Verifying the identity of the person
2. Validating the person accessing the service is the same person who performed step 1.

 

 

 

In the original version of the identity verification process, the end user would produce acceptable ID documents and the verifier would check both the validity of this document and then the fact that the same person was applying for the service. It became clearer that this entire process would need to be substituted by a digital solution that substituted the verifier’s ability to perform both these actions.

Technology fitment:

An IDVT solution would be best suited as understood from the previous phase. We thus systematically began speaking to various technology vendors to help our customer partner choose the best solution that had the following elements in place

1. A document authentication via scan
2. Extraction of document data via OCR with negligible loss of information for accurately recording customer data.
3. Native mobile SDKs as user research indicated this was most likely to comfortably work in a mobile environment with phone cameras.
4. Face-match technology to validate the user applying is the same as recognized in a passed document check
5. Validation that 2D images were not used to bypass face-match technology.
6. Vendor with a large document base with experts in the field of document technology to give expert supportive opinion as a fallback for suspicious cases of document fraud.

We settled eventually to integrate with a prominent IDV partner in the UK and incorporated their mobile SDKs in our existing native Android and iOS apps. An in-depth understanding of the problem space and now the solution space, we were able to leverage all aspects of the technology available like fine-tuning face-match thresholds, ensuring a live person was applying by performing real time expressions to prove their actual presence when applying among many others. All this, for the end user from the comfort of their homes on their mobile phone.

Post development:

Our customer went on to sell 100k+ of these ID backed products. We also helped tailor some of their customer support process to the auditory and compliance requirements that the framework required to ensure customer data was accurately recorded and maintained securely. The solution also needed to be GDPR compliant.
A few months down the line, the same IDV concept was incorporated into a greenfield product licensed as a B2B service to the customer’s partners for onboarding the latter’s customers in the peak of the pandemic.

Based on the above experiences we have had, here are some tips for implementing your identity solution

1. UX is still King!

Just leveraging digital ID tech is not enough. The instructions to users and messaging throughout the process needs to be as consistent. As a thumb rule, consider that a good 80 percent of your user base will have never attempted something similar. In our case, we had not just complexities of face-match and verification to contend with but also some aspects of quality of photograph that would eventually be needed for the service. One has to very unambiguously inform users on what aspects can go wrong and what they need to do next. Drop-offs are an even bigger risk with ID verification as it is still novel in many parts of the world.

2. Pay attention to on demand cost of verification

Most services will charge on demand on a cost per verification basis. Work out bulk propositions with vendors and ensure that they are constantly improving quality of verifications so you are paying only for the valid verification attempts. Again a clear and consistent user experience helps to ensure users aren’t struggling continuously and burning through your total accounted-for service cost. Limit number of user attempts before they need help for support intervention. Issues can very often be down to users not performing instructions correctly, limited support on older devices or an odd identity document format that has not been added yet to the vendor databank of verified document formats.

3. Continuously improve

Things are rapidly improving in this space and your vendor may be making improvements to their SDKs. Make sure you stay abreast and watch out for anything that reduces customer friction.

4. Understand limitations

It’s important to be transparent about the limitation of image recognition technologies and potential pitfalls it may bring. Face-match and AI algorithms are constantly evolving and often start off working on limited datasets and ethnicities. It’s important to recognize the impacts to your business early and make provisions. E.g. We were using a human face recognizing algorithm from a famous cloud provider prior to the identity verification to ensure users didn’t mistakenly upload their ID card photos as the actual photograph or photos of their pets! We made sure that the threshold of acceptance was lowered simply because the AI could mess up on the face recognition when provided images of users from a non-Caucasian ethnic origin. Image if 50 % of your user base was an international audience!

With the right expertise backing your proposition, there is indeed a lot of value in making effective use of digital ID tech to improve operational efficiencies in your onboarding processes, save time and resource and prevent fraud better and faster than what human eyes might be able to detect. Even if compliance requires you to still have manual checks in place. IDV helps speed the process regardless by aiding decision making.

For more information on how we can help you on leveraging Digital Identity for your value proposition, contact us at talktous@arrkgroup.com