Banking & capital markets
Compliance evidence at trade speed
Pre-trade MiFID II checks and DORA evidence on demand, inside your perimeter.
AIA · Agentic Intelligence Architecture
AI agents for
regulated industries
Open-source AI agent infrastructure for regulated enterprises.
Financial services
Healthcare & life sciences
Government & defence
Manufacturing & supply chain
The problem
Frontier AI ships your records to a vendor's cloud — someone else's servers, someone else's jurisdiction.
With AIA, data stays in your perimeter — and your auditor can verify what the agents did.
Industries
Built for regulated industries
Healthcare & life sciences
Analytics without PHI egress
Federated cohort analytics with zero PHI egress, 21 CFR Part 11 end to end.
Government & defence
Classified work, on your terms
Classified synthesis on air-gapped hardware, sources and methods protected.
Legal & compliance
An audit trail any party can replay
Privilege-preserving e-discovery and contract mapping with reproducible provenance.
Regulations
Aligned to the frameworks
Aligned to the frameworks
that govern your industry
DORA
Digital Operational Resilience Act · live since Jan 2025
EU · Financial services
EU AI Act
High-risk obligations · Art. 12 logging · 2 Dec 2027
EU · Cross-sector
NSM-10
Federal post-quantum migration
US · Federal
BSI TR-02102
PQC migration guidance
Germany · Cross-sector
Architecture
Three pillars
The system to win the agentic-AI era.
Pillar 01
Multi-agent orchestration
Specialised agents across security, finance, and legal, each task routed to the best fit.
Pillar 02
Knowledge graph
A Neo4j-backed graph where answers trace to the exact nodes behind them.
Pillar 03
Post-quantum A2A
Post-quantum agent-to-agent protocol. Actions are signed into a tamper-evident log, designed to anchor a single zero-knowledge proof on-chain.
How AIA works
Run the agents
Run the agents
where your data already lives
Agents read, reason, and work inside your perimeter; all that leaves is a zero-knowledge proof.
Five stages inside the perimeter. One proof leaves.
The AI audit layer
One audit layer.
One audit layer.
Three kinds of proof.
The audit trail that gets your AI agents approved. AIA writes three kinds of proof; each shows that something is true and nothing else.
Action attestation
Actions are signed into a tamper-evident log. Its root is designed to anchor on-chain, so an auditor can confirm an action happened and was not altered — without seeing the data behind it.
Confidential transfer
Payments between agents and services settle in regulated stablecoins. Amounts and counterparties stay private, disclosed only to a party you choose.
Verifiable inference
For self-hosted models: hardware attestation — a confidential-GPU receipt binds the output to the model and policy. For frontier APIs: every call is attested — request, response, model identity as reported. Cryptographic proof of inference (zkML) is our research north-star.
Proofs are batched, so anchoring thousands of actions costs a single settlement. Run it in our EU regions, in your VPC, or fully air-gapped — same code, same proofs.
Architecture
| Proof | What it proves | Aligned to |
|---|---|---|
| Action attestation | An agent did a specific action, unaltered, at a known time. | EU AI Act Art. 12 · DORA · NIS2 · HIPAA audit trail |
| Confidential transfer | Value moved between parties, amounts and counterparties private. | MiCA · regulated stablecoins · selective disclosure |
| Verifiable inference | An output came from an approved model under an approved policy. | EU AI Act high-risk obligations |
Deployment
From managed cloud
From managed cloud
to air-gapped on-premise
Three patterns, one codebase. Identical APIs and proofs across all three.
We operate it
013a runs the cluster; you hold the keys.
Inside your cloud
Runtime, graph, and keys inside your VPC; only proofs cross the boundary.
Air-gapped
Full on-premise install with no outbound connectivity, for classified workloads.
Who builds it
Built in Berlin by 013a
Deep expertise in cryptography, multi-agent systems, and enterprise security — from the EU, for the regulated world.
Made in Germany
Start
Two paths in one conversation
A thirty-minute call to map your workload to the right deployment, agents, and audit evidence. No deck.