Premium Drop3 Startups · $25M · AI Verification + Semantic Engineering + Travel AI
2 startups with founder intel, hiring signals, and outreach playbooks.
Ciphero
ciphero.ai · New York, NY · AI Verification Layer for Enterprise
What they're building
Ciphero builds an AI verification layer for enterprise environments, a system that captures, checks, and governs every AI interaction that occurs inside an organisation, whether human-initiated or agentic. As enterprise AI adoption has accelerated, the number of interactions with tools like ChatGPT, Copilot, Claude, and internal AI agents has exploded, and most security teams have no visibility into what data those interactions are touching or what those models are returning. Ciphero installs a sovereign control layer across all of it: auditing inputs and outputs in real time, flagging policy violations, detecting prompt injection attacks, verifying outputs against internal truth sources, and surfacing shadow AI usage that legacy DLP tools and static firewalls cannot see. The platform tracks behaviour across employees, models, and agent-to-agent interactions simultaneously, and enforces governance policies dynamically rather than through static rules. The pricing model is flat, not usage-based, which removes the friction that causes enterprise security buyers to throttle adoption.
Why this matters
The numbers the company cites are not marketing estimates. They are the kind of figures that land in a CISO's quarterly review: 80% of enterprises show signs of shadow AI activity, one in six data breaches this year was traced to unauthorised AI usage, and 97% of organisations have no proper AI access controls in place. Enterprise AI governance barely existed as a compliance category eighteen months ago. Now it is in the top three items on every CISO's agenda in financial services, healthcare, legal, and government. Ciphero's founders built Fakespot, a detection and verification system for fake reviews and counterfeit listings, which Mozilla acquired in 2023. The underlying capability - identifying when an AI-generated or AI-mediated output is not what it claims to be - is exactly what Ciphero applies to enterprise AI interactions. This is not a pivot. It is the same detection architecture applied to a much larger and faster-growing threat surface.