Dynamic Fallback Workflows for Mobile Liveness Detection

Optimize User ExperienceImplement multi-layered liveness detection with dynamic fallback options to reduce friction and increase successful verifications, especially in challenging environments like low light or with older devices.

Enhance Fraud PreventionLeverage a combination of Passive, 3D Flash, and 3D Action & Flash liveness methods, dynamically switching to more robust options when initial attempts indicate higher risk or potential spoofing.

Maintain Security StandardsConfigure thresholds and actions for various risk factors—such as low liveness scores or potential duplicate faces—to automatically trigger review or decline, ensuring compliance and preventing sophisticated attacks.

Didit's Modular AdvantageDidit's AI-native platform allows businesses to easily build and orchestrate flexible liveness workflows with configurable rules, offering Free Core KYC and no setup fees for scalable, secure identity verification.

In today's digital-first world, biometric liveness detection has become a cornerstone of secure identity verification in mobile applications. From banking and fintech to social media and e-commerce, ensuring that a real, live person is interacting with the app—rather than a fraudster using a photo, video, or deepfake—is paramount. However, relying on a single liveness detection method can lead to a suboptimal user experience, with legitimate users potentially failing verification due to factors like poor lighting, older devices, or even a simple blink at the wrong moment. This is where a dynamic fallback workflow becomes indispensable.

A dynamic fallback workflow intelligently adapts the liveness detection process based on initial results, user environment, and configured risk thresholds. Instead of a one-size-fits-all approach, it allows for a seamless transition between different liveness methods, optimizing for both security and user pass rates. Didit, with its advanced Passive & Active Liveness capabilities, provides the ideal foundation for building such resilient systems.

The Need for Dynamic Liveness Workflows

Traditional liveness detection often presents a binary outcome: pass or fail. While effective for basic spoofing attempts, this rigid approach can lead to user frustration and abandonment. Imagine a user attempting to open a new bank account, only to be declined because their phone's camera quality isn't optimal for the required liveness check, or they're in a dimly lit room. A dynamic workflow addresses these challenges by offering alternative paths to verification without compromising security.

For instance, an initial attempt might use a less intrusive method, like Didit's Passive Liveness, which analyzes a single frame for signs of liveness. If this fails due to low face quality or a potential spoofing attempt, the system can automatically prompt the user to try a more robust method, such as 3D Flash. This escalation ensures that legitimate users can still complete their verification while fraudsters are met with increasingly sophisticated defenses.

Moreover, the threat landscape is constantly evolving. Deepfakes and advanced presentation attacks demand a flexible defense. By dynamically adjusting the liveness method, businesses can stay ahead of fraudsters, deploying the appropriate level of security for each unique situation. Didit's AI-native approach ensures that its liveness detection methods are continuously learning and adapting to new attack vectors.

Designing Your Dynamic Fallback Strategy

Building an effective dynamic fallback workflow involves several key considerations:

1. Prioritize User Experience: Start with the least intrusive and fastest method. Didit's Passive Liveness is excellent for this, offering standard security with minimal user effort. If the score is high and no warnings are triggered, the user can proceed quickly.

2. Define Risk Thresholds: Set clear thresholds for liveness scores and warning types. Didit's Liveness Detection report provides detailed scores and warnings (e.g., LOW_LIVENESS_SCORE, LIVENESS_FACE_ATTACK, POSSIBLE_DUPLICATED_FACE). You can configure your application to automatically approve, review, or decline based on these. For example, a score below 70 might trigger a fallback, while a score below 50 might lead to an automatic decline.

3. Implement Tiered Liveness Methods:

   • Tier 1 (Standard Security): Passive Liveness. Fast, convenient, and suitable for low-risk use cases or as a first attempt. If a LOW_LIVENESS_SCORE warning is triggered, or if MULTIPLE_FACES_DETECTED (in passive mode) or LOW_FACE_QUALITY occurs, move to Tier 2.
   • Tier 2 (High Security): 3D Flash. If Passive Liveness indicates a higher risk or fails, prompt the user for a 3D Flash check. This method projects dynamic light patterns to create a depth map, providing high security against photos or 2D spoofs while maintaining a seamless experience.
   • Tier 3 (Highest Security): 3D Action & Flash. For the highest-risk scenarios (e.g., high-value transactions, account recovery), or if 3D Flash still raises concerns (e.g., LIVENESS_FACE_ATTACK), escalate to 3D Action & Flash. This combines randomized actions (like blinking) with dynamic light analysis, making it nearly impossible to spoof with deepfakes or advanced masks.

4. Handle Automatic Decline Conditions: Certain conditions should always result in an automatic decline, regardless of the fallback strategy. Didit flags these clearly, such as NO_FACE_DETECTED, LIVENESS_FACE_ATTACK, or FACE_IN_BLOCKLIST (if the face matches an entry in your blocklist via Didit's 1:1 Face Match & Face Search). These are non-negotiable security failures.

5. Provide Clear User Guidance: When a fallback is triggered, clearly explain to the user why a different step is needed and how to successfully complete it. This reduces frustration and improves completion rates.

Leveraging Didit's Liveness Detection Report for Orchestration

Didit's comprehensive Liveness Detection report is key to building these dynamic workflows. Returned as a JSON object, it provides critical information:

• status: Overall verification status ('Approved', 'Declined', 'In Review', 'Not Finished').
• method: The specific liveness method used ('ACTIVE_3D', 'FLASHING', 'PASSIVE').
• score: A confidence score indicating the likelihood of liveness.
• age_estimation: Useful for age-restricted services, directly integrated within the liveness response.
• warnings: A crucial array detailing any risks detected, such as LOW_LIVENESS_SCORE, POSSIBLE_DUPLICATED_FACE, or HIGH_FACE_LUMINANCE. Each warning includes a risk type, short_description, and log_type.
• matches: If Face Search is used, this shows matching sessions and their similarity percentages, including whether a matched face is is_blocklisted.

By parsing this report, your application can make intelligent, real-time decisions. For instance, if the score is below your 'review threshold' but above your 'decline threshold', and a LOW_LIVENESS_SCORE warning is present, your workflow can automatically initiate a prompt for a higher-security liveness check using Didit's 3D Flash method. If FACE_IN_BLOCKLIST is detected, the transaction can be immediately declined.

How Didit Helps

Didit is uniquely positioned to help businesses build robust and dynamic fallback workflows for biometric liveness detection. Our modular, AI-native platform offers a comprehensive suite of tools designed for flexibility and security:

• Modular Liveness Methods: Didit provides Passive, 3D Flash, and 3D Action & Flash Liveness, allowing you to choose and dynamically switch between methods based on your specific security needs and risk appetite. This modularity means you can start with a basic check and escalate as required, ensuring both security and user experience.
• AI-Native Accuracy: Our liveness detection boasts 99.9% accuracy with a false acceptance rate of less than 0.1%, effectively combating sophisticated spoofing attacks including deepfakes.
• Configurable Workflows: With Didit's no-code Business Console and clean APIs, businesses can easily orchestrate complex identity verification workflows. You can define rules, set thresholds for liveness scores, and configure actions for various warnings (e.g., 'Review' for LOW_FACE_QUALITY, 'Decline' for LIVENESS_FACE_ATTACK).
• Comprehensive Reporting: The detailed Liveness Detection Report provides all the necessary data points—scores, methods, warnings, and metadata—to power your dynamic decision-making engine.
• Developer-First Approach: Didit offers an instant sandbox and public documentation, making it easy for developers to integrate and customize liveness detection into their mobile applications.
• Free Core KYC: Get started with essential identity verification features at no cost, allowing you to implement robust liveness detection and explore dynamic workflows without initial investment.

By leveraging Didit's capabilities, businesses can create a seamless yet secure user journey, reducing friction for legitimate users while effectively deterring fraudsters. The flexibility of Didit's architecture ensures that your liveness detection strategy can evolve with the ever-changing threat landscape.

Ready to Get Started?

Ready to see Didit in action? Get a free demo today.

Start verifying identities for free with Didit's free tier.