resistant ai logo

Resistant AI

Resistant AI protects the automation and AI systems of financial services from manipulation and attack. It subjects every customer interaction — from documents submitted at onboarding to ongoing behaviors — to forensic analysis to detect document forgery, serial fraud, synthetic identities, bots, account takeovers, money laundering, and unknown financial threats operating at scale.

Location:Prague, Czech Republic
Founded: 2019

Fraud Solution Profile

Criminals are iterating at the speed of startups, finding new weaknesses in the AI and automations of financial services, and using automations of their own to exploit them at scale. The only way to tackle these unknown, emergent threats is to submit every customer interaction—from the documents they submit to the transactions they perform and the behaviors they exhibit— to a level of scrutiny previously only possible to human review, but at scale and economically. 

Resistant AI provides that scrutiny by layering on top of existing customer systems. It uses data from across the entire customer journey, augmenting all existing tools and the human teams who use them with truly smart AI that protects them from manipulation and attack. 

Resistant AI has 2 main products which can be combined to create a feedback loop, and can leverage the data from existing systems: 

Resistant AI Document Forensics

While document automation is desirable, doing so without regards to fraud is dangerous. Most document fraud today is invisible to the human eye and easy to mass produce to overwhelm any fraud, risk, or compliance team. Resistant AI’s Document Forensics is the leading document fraud detection service used by leading financial institutions and powering many ID and third party verification services. It analyzes any digital document — bank statements, payslips, tax forms, business registrations, ID cards, and more submitted in PDF or image formats — over 500 different ways to detect signs of forgery. It then compares all submitted documents to find forgery patterns across applicants to uncover document reuse, template farms, and mass fraud attempts. Finally, session data from the onboarding process (device fingerprints, timestamps of actions, etc) and PII can be analyzed to highlight behavioral anomalies and identity overlaps to stop serial onboarding attempts.


Clear verdicts (not scores) are provided with supporting analysis in human readable language that have been used in court cases. Both verdicts and individual indicators can be used to create hyper-granular triggers for acceptance, decline, or escalation workflows. 


Customers have seen: 

  • 30x in ROI in automation savings alone
  • Manual reviews reduced by 92%
  • Application review processes accelerated by up to 80%
  • Confidence returns to investigators who can now strategically focus on fraud trends.


Resistant AI Transaction Forensics

Resistant AI Transaction Forensics is a practical application of AI that fixes existing Fraud and AML transaction monitoring systems and boosts investigator productivity. It simplifies the rules used to detect fraud and financial crime to surface fewer but higher quality alerts for review, highlighting previously hidden behaviors using layers of modular, ready-built ensembles of models that keep deployments fast and maintenance low. 


By combining dynamic behavioral segmentation with identity clustering techniques, Transaction Forensics:

  • Reduces the total rulesets in the underlying system  that need to be updated and maintained, 
  • Reduces the false positive noise by up to 95%
  • Increases findings of new, unknown risks by over 200%
  • Categorizes the alerts for greater investigation productivity
  • Limits their exposure to fraud and financial crime  with pre-transaction monitoring capable of operating under 50ms.




Raiffeisen Bank

Woodside Credit

Ecommerce, Financial Services, Insurance
 Primary Functionality
Fraud & AML Platform
 Fraud Type
Account Takeover, KYC & AML, Loyalty or Promo Abuse, New Account Fraud, Payment Fraud, Synthetic Identity Fraud
Behavioral Biometrics, Machine Learning