Investment Opportunity Β· 2026

Funding the Future of
Human–AI Epistemic Safety

Kermati is building the world's first formally grounded middleware architecture for human-AI trust calibration β€” across three structured funding phases.

$5,000,000 Total Raise Β· 3 Phases

Investment Structure

Three Phases. One Architecture.

Development and growth is structured across three phases β€” each building directly on the prior, from core infrastructure validation to full-scale deployment in high-risk professional sectors.

Phase 1 - Circuit Core

Phase 01

Operational Launch

$250,000

Epoche v3.0 EOF Pipeline

The Pre-Seed round. This is our most critical phase β€” it funds the team and infrastructure needed to validate core epistemic governance middleware. Epoche's hallucination-risk detection and the EOF multi-agent pipeline are built and stress-tested here.

⬑ 150-person pilot lab · Initial validation cohort
Phase 2 - Growth

Phase 02

Empirical Validation

$1,750,000

Epoche v3.0 EOF Pipeline Epoche Bridge

Architecture hardening and EU AI Act compliance validation. We shift into full velocity β€” meeting regulatory deadlines, expanding the pipeline infrastructure, and proving the Epoche–EOF stack against real-world adversarial data in controlled environments.

⬑ 150-person pilot lab · Active validation cohort
Phase 3 - Lotus

Phase 03

Full-Scale Deployment

$3,000,000

EAS Framework Sozo Suite

This is when we show the world how deep the roots go. The Epistemic Agency Suite and Sozo integration require a 1,500-person research lab drawn exclusively from high-risk professional sectors β€” the populations where AI-assisted decision failure carries the highest human cost.

⬑ 1,500-person research lab · High-risk sector professionals

Phase 3 Research Cohort

High-Risk Sector Professionals

The EAS and Sozo validation research requires 1,500 participants sourced exclusively from domains where AI-assisted decision failure carries direct human consequences.

πŸ₯

Medicine

Physicians, nurses, diagnosticians β€” high-stakes clinical AI use

βš–οΈ

Law

Attorneys, judges, paralegals β€” AI-assisted legal research and reasoning

βš™οΈ

Engineering

Systems and structural engineers β€” safety-critical technical decisions

πŸ“ˆ

Fintech

Analysts, traders, risk managers β€” high-velocity AI-augmented decisions

πŸ”¬

Research

Scientists and policy experts β€” epistemic integrity under automation

Why Now

The Investment Thesis

Two co-occurring AI failure modes β€” hallucination and automation bias β€” create a compounding risk that neither AI safety nor UX research has formally addressed together.

The Problem

AI hallucination and automation bias are not independent failures β€” they are a coupled system. High-confidence incorrect AI output accepted without scrutiny compounds undetected. No current product formally addresses both simultaneously.

The Architecture

Epoche provides interaction-layer friction. EOF provides multi-agent consensus. EAS quantifies epistemic independence in real time. Sozo integrates behavioral science to make friction motivationally effective, not just technically correct.

Regulatory Tailwinds

The EU AI Act's high-risk provisions require algorithmic traceability, audit trails, and human oversight mechanisms that the Kermati stack satisfies by design β€” positioning Epoche as compliance infrastructure, not a feature.

Defensible Moat

The mathematical substrate β€” HII, CVD, Ξ², EFC, HRI β€” is formally specified and peer-grounded. The behavioral layer is built on published dynamic norm theory. This is not a prompt wrapper. It is institutional-grade governance middleware.

Ready to be part of this?

Contact the Kermati team to receive our full investor package and projected financials.

Request Investor Package