How Signicat uses AI to prevent identity fraud and build digital trust

AI is changing digital identity on both sides.

Fraudsters use it to create deepfakes, synthetic identities, forged documents, personalised phishing and automated attacks. Signicat’s 2025 identity fraud report revealed that 71% of firms agree that AI is behind most of the fraud they face.

So for Signicat, AI is not a buzzword. It is a practical tool for preventing identity fraud and building trust.

Signicat uses AI where it creates measurable trust: helping verify real users, detect fraud signals, reduce friction and support smarter risk decisions across digital identity journeys. The goal is simple: help genuine users move faster, while making it harder for fraudsters to get through.

AI-powered video verification for real-time fraud detection

One of the clearest examples is VideoID, Signicat’s AI-powered video identity verification solution.

VideoID uses AI and machine learning for real-time document and biometric checks. It helps detect fraud signals such as deepfake impersonation, presentation attacks and injection attacks. Each video frame is processed by at least 10 machine learning models, including likeness, liveness, Presentation Attack Detection and Injection Attack Detection models. Signicat also states that more than 1 million videos are processed annually using advanced machine learning.

That matters because digital identity fraud is becoming more realistic. A human reviewer can miss subtle manipulation. A static rule can fail when the fraud pattern changes. AI helps detect the small signals that are difficult to spot at scale.

Securing logins and high-risk transactions with AI face authentication

Fraud does not stop after onboarding. In many cases, the real risk appears later: account takeover, payment fraud, eID misuse, phishing or account recovery abuse.

That is where MobileID Face Authentication comes in.

Signicat describes Face Authentication as a MobileID feature that uses AI and machine learning for fraud prevention during login and authentication. It uses a 2-second facial scan, advanced biometrics, 3D liveness checks and facial mapping, performed in the customer’s mobile app and verified server-side.

The result is stronger protection when it matters most: login, recovery, payment approval and other high-risk actions.

The role of AI in secure qualified electronic signatures (QES)

AI also plays a role when identity assurance and legal certainty meet.

With SignatureID QES / QESMulti, Signicat combines qualified electronic signatures with video-based identification. The onboarding process is analysed by AI, recorded and checked by a qualified agent. The solution also uses deep learning biometrics, AI-assisted document recognition and OCR data extraction with more than 20 real-time verification and authentication checks.

For users, this creates a digital signing journey they can complete remotely. For organisations, it supports compliance, security and conversion in high-assurance signing flows.

Using machine learning to detect hidden fraud in large datasets

Some fraud is obvious. Much of it is hidden in patterns.

Signicat has worked with machine learning to detect fraud and anomalies in large authentication datasets. One example is the use of K-means clustering in BigQuery ML to identify unusual behaviour based on signals such as geolocation, device information, login time and authentication method.

A single unusual login may not prove fraud. But a login from a new location, at a strange time, from a different device, using a different method, can become a signal worth investigating.
This is a good reminder that AI is not only generative AI. Sometimes the right tool is machine learning, anomaly detection and well-prepared data.

Building complex identity workflows with a no-code AI assistant

Not all AI is about stopping fraud. Some of it is about helping teams move faster.

MintyAI, an AI assistant for Signicat Mint – the no-code digital workflow builder – helps users build identity workflows using natural language. A user can describe the flow they want, and MintyAI can suggest a draft workflow for identity proofing, authentication or electronic signing. It can also explain workflow steps and turn an AI-generated response into a functional workflow inside Signicat Mint.

This makes complex identity journeys easier to start, understand and improve. Human experts still make the final decisions, but AI helps remove friction from the building process.

Boosting internal productivity with enterprise AI platforms

AI also supports internal productivity at Signicat.

With tools such as Gemini Enterprise, teams can search across business data, connect workplace tools, automate multi-step workflows and build AI agents for specific tasks. Gemini Enterprise as an agentic platform that connects to tools such as Microsoft 365, OneDrive, SharePoint, HubSpot and Jira, while supporting governance, permissions and security controls.

For Signicat, this means AI can help reduce repetitive work, improve access to knowledge and support better decision-making across teams.

The same applies to documentation and support. Signicat’s support page points users to an AI assistant in the developer documentation and Dokobit help centre, helping users find answers faster through natural-language questions.

Responsible AI: Ethically training models to combat identity fraud

AI models need to improve as fraud evolves. But in digital identity, that must be done carefully.

Signicat’s biometric data collection notice explains that, with consent, video identification data may be used to generate training datasets to improve AI applied to digital images. The purpose is to improve detection of document forgery and identity fraud, including forged documents, face impersonation, manipulated images and related attacks.

That is the principle behind how Signicat uses AI: practical, layered and responsible.

Fraudsters use AI to scale deception.

We use AI to scale trust.