Variance is replacing traditional, human-led risk and compliance workflows with purpose-built AI agents. These agents can reason over unstructured data, access internal and external data sources (including the open web and legacy UIs), and apply company-specific policies to automate 99% of cases, from content moderation to complex financial verifications.
Variance intentionally operated in stealth for three years, building its product and customer base quietly. The company positions itself as a 'secret weapon' for its clients, as publicizing its methods could tip off bad actors and create new vectors for fraud and abuse.
The company achieves the software output of a much larger organization with a small engineering team of five. This is accomplished by deeply integrating AI coding agents into their workflow, where each engineer effectively manages a small team of AI assistants to accelerate development.
Variance's platform goes beyond typical fraud to identify and investigate severe threats with real-world consequences. The AI agents' ability to analyze relationships between entities at scale has enabled the detection of complex fraud rings, state-sponsored actors pushing specific narratives, and credible online threats of physical harm.
Keep pulling the thread on Karine Mulata.