Founder
Hai Cao
Security engineer with a background in security operations and cloud security. Building Verdikor to make AI in the SOC verifiable.
Company
We're building the check that AI in security has been missing.
When a CISO or an auditor asks whether the AI in a SOC can be trusted, we want the standard answer to be: “We run it through Verdikor. Here is the verification report.”
Why we exist
The person grading the work shouldn't be the one who did it.
Every SOC is buying AI to investigate faster, and almost no one lets it act alone. Each vendor grades its own AI. Nobody independent checks what the AI said against the customer's real data, across every SIEM and every AI.
That check can't be another model. Models that review their own kind share the same blind spots. It has to be deterministic, tied to raw evidence, and independent of whoever did the investigation.
That is Verdikor: one independent check for every AI and every SIEM a SOC runs.
Commitments
What we will and won't do.
We will
- Publish the evidence format and a free verifier
- Show numbers measured on real cases before we ask anyone to believe them
- Support every major SIEM, with no favourite
We won't
- Build an AI analyst. The day we investigate, we stop being independent.
- Put an LLM in the decision path
- Lock your knowledge into our format. It lives in your git.
Founder
Built by an engineer who has wired AI into a SOC.
Want to work on this? Write to hello@verdikor.com.
Early access
Run Verdikor on your own cases.
We're working with 3 to 5 design partners: MSSPs running several SIEMs, and banks or insurers using AI with any SIEM.