Deutsche Bank deploys agentic AI to cut vendor risk assessment time from 30 minutes to under two minutes

Deutsche Bank's TPRM AI cuts vendor risk document review from 30 minutes to under two minutes, hitting 90% accuracy. Human assessors retain final decision authority.

Categorized in: AI News Management
Published on: May 15, 2026
Deutsche Bank deploys agentic AI to cut vendor risk assessment time from 30 minutes to under two minutes

Deutsche Bank Uses AI Agents to Speed Vendor Risk Assessment

Deutsche Bank's Global Procurement & Vendor Management team deployed an AI system in December 2025 that cuts third-party risk assessment time from 30 minutes per document to under two minutes for up to five documents. The system, called TPRM AI, uses multiple specialized AI agents to analyze vendor evidence and propose outcomes, with human assessors making final decisions.

Vendor onboarding at Deutsche Bank requires teams to verify that external partners meet internal standards and regulatory expectations. Before vendors can work with the bank, they must submit documentary evidence that assessors check against a control framework. The process demands expertise, careful judgment, and significant time investment.

How the system works

TPRM AI coordinates several AI agents that work sequentially. One agent retrieves the relevant Deutsche Bank control questions. Another drafts answers based on vendor evidence. A third proposes an outcome and cites the exact source passages. Human assessors then review these suggestions and make the final decision.

The speed gains compound in real-world scenarios. When vendors submit business continuity evidence across four or five separate files, manual review typically takes three hours. With AI-supported document analysis followed by human validation, the same work now takes about 30 minutes.

Accuracy and oversight

TPRM AI achieves around 90% accuracy in recommended outcomes when benchmarked against human validation, with accuracy expected to improve as the system evolves. Every suggestion includes citations to source passages, making verification transparent and keeping final decision rights with trained human assessors.

The cloud-based design allows the system to scale as assessment volumes grow. The bank sees potential for significant cost avoidance with a full rollout across vendor management functions.

Broader implications

The deployment represents a first milestone in applying AI agents and automation to third-party risk assessment. The foundation created here could extend to other risk assessment areas within the bank, improving internal efficiency and client service without proportional increases in staffing.

For management professionals overseeing vendor relationships or risk assessment functions, understanding how AI can support rather than replace human judgment in compliance work is becoming essential. The AI Learning Path for Management Consultants covers how to evaluate and implement similar systems in your organization.


Get Daily AI News

Your membership also unlocks:

700+ AI Courses
700+ Certifications
Personalized AI Learning Plan
6500+ AI Tools (no Ads)
Daily AI News by job industry (no Ads)