Professional Services|It|Reduce Risk

KPMGAI application security

Daily GenAI use introduced prompt injection risks. Real-time monitoring now blocks threats and addresses 8 of 10 top LLM security risks.

Nov 2, 2025|3 months ago

Key results

LLM Security Risks Addressed
8/10

The company

Global professional services network for audit, tax, and advisory solutions.

IndustryProfessional Services
LocationAmstelveen, Netherlands
Employees100K+
Founded1987

Result highlights

  • 8 of 10 OWASP LLM security risks addressed

The story

A "Big Four" financial audit and advisory firm with 40,000 employees operating across 95 locations in the United States, Israel, Mexico, and Latin America.

Daily employee usage of generative AI introduced critical risks, including prompt injections and data poisoning identified in the OWASP Top 10 vulnerabilities. The firm needed to safeguard model lifecycles and govern sensitive data across its complex multi-tenant environment.

The firm deployed Microsoft Defender for Cloud to monitor Azure OpenAI Service apps and detect threats like jailbreak attempts in real time. Microsoft Purview automatically scans and applies sensitivity labels to classify data, enforcing "secure by default" policies across the organization. Endpoint agents on developer laptops monitor prompts for risks, feeding signals to Microsoft Sentinel for automated remediation.

Quotes

Explore similar

Find AI opportunities for your
business context

Understand what's working with 2,383 recent AI case studies across industries. We structure things so you can find high-impact strategies for your exact context.

Graphic placeholder

Industries covered