Sber
Face and voice ID module
2019
Enhanced financial security through user-friendly face and voice ID functionality in the mobile app of Russia's largest bank, Sberbank.
That was an activity within the framework of a government initiative to implement unified biometric identification across all financial organizations in the country.
As part of this program, Sber implemented its own face and voice ID solution and integrated it across all bank products.
The challenge was to implement the new module into the existing functionalities, customer scenarios, and business processes seamlessly and smoothly.
Additionally, I needed to ensure it did not disrupt the habits of diverse user groups, such as older adults or clients with disabilities.
My role was to design the simplest possible solution for users while considering technical and business requirements.
The module comprises multiple layers of identification, which can be invoked in various combinations by a compliance and security module based on a client's activities and the monetary amount involved in a specific transaction. The ID module is integrated into the security system workflow and functions within the app when necessary.
Following successful identification, the module verifies the user, and the process proceeds to completion, such as initiating a transfer or signing an online application.
I closely collaborated with the tech lead and business analyst to effectively align technical features with business goals while maintaining a user-oriented focus.
Leading this communication enabled me to define the details and, through several iterations, ultimately provide a polished user flow.
This helped reduce friction for users and gave developers a better understanding of how the functionality integrates with existing modules.
Given that such functionality involved using device modules and sensors, it was impossible to test it with prototypes in Sketch. Instead, we built a simple solution with a real identification library and tested it through several iterations within the team, the company, and with bank clients.
Each iteration helped us improve the accuracy and speed of interaction with the interface.
For example, the module required brief processing once basic facial points, such as eyes and gaze direction, nose and ears, and mouth position, were recognized.
Therefore, we implemented animations and vibration feedback to inform the user that their face was captured.
The app opens the camera interface and prompts users to align their face within the frame.
The system dynamically tracks the user's face, guiding them to the correct position if they turn their face to the side, up, or down. Text notes are displayed for additional assistance.
For added security, such as during high-value transactions, users may be prompted to verbally provide several digits.
The app gives clear instructions for the user to follow specific conditions.
If identification fails, it shows reasons and suggests ways to fix it. The user gets a few tries, and if it still doesn't work, they can confirm the operation in other ways.
All disabled clients have specific markers in their profiles, enabling the provision of even more personalized service across all bank products and channels.
For users with vision disabilities, the app offers assistance through voice-over functionality and vibration feedback.
The UI elements are a bit more contrasted and larger.
Following a principle similar to a car parking aid, the vibration increases as users align their face more accurately, complemented by voice assistance.
The functionality enables customers to confirm specific operations in a matter of seconds, not minutes, significantly reducing the call center’s load.