Mike
Founder & Funder
Leads strategic direction and funding. Experience in public-sector transformation, compliance and institutional technology adoption.
Explainable Decision Support
Explainable decision infrastructure for regulated teams
Turn policies, legal criteria and internal rules into reproducible decisions with audit trails and human oversight.
Built for HR, compliance, legal, and risk teams in public institutions, banks, insurers, and law firms.
Same inputs → same outputs
Complete audit logs
No hidden logic – fully explainable
Configurable fairness checks
Nominated – European Ombudsman Award 2025–2026
Designed to support governance, transparency and human oversight requirements
CASE: MOB-2025-03 // MOBILITY ALLOWANCE
⚡ RULES v2.4
🔒 AUDIT ID: AUDIT-2025-03-15-001
CONTRACT DETAILS • REF: EMP-4281
Human oversight
Audit-ready
On-prem / VPC
DPA-ready
Air-gap compatible
How we combine deterministic and generative
A layered approach: generative for summarisation, deterministic for rules, human for approval.
⚖️
Deterministic Rule Engine
Same inputs → same outputs. Every decision is traceable to a policy ID. No hallucinations, no probabilities.
Policy ID MOB-2025-03
Version v2.4.1
Audit hash a3f2…
🧠
Generative AI (LLM) for drafting
Purely generative black-box outputs can be difficult to justify, audit and operationalise in regulated decision workflows. We use generative AI only for summarisation and draft narratives — never for final decisions.
📝 summary only
👤 human review
🔍 full audit
Governance Pipeline
Unidirectional, deterministic, fully auditable. Every step is logged and versioned.
1. Ingestion & Normalization
2. Deterministic Rule Engine
Maps legislation to logic. Built for predictability.
3. Reasoning Trace Generation
4. Human-in-the-loop Gate
Mandatory officer sign-off
5. Immutable Audit Log
System Components
Modular, API-first, deployable in air-gapped environments.
RULEBASE
Structured Reasoning Engine
Deterministic, transparent, version-controlled. Predictable outputs for policy execution.
AUDIT
Immutable Audit Trail
Append-only logs, cryptographically signed. Tamper-evident by design.
HITL
Human-in-the-loop Gate
Mandatory review and override with justification. Designed to support compliance governance.
FAIRNESS
Embedded Fairness Checks
Reproducible checks at decision points. Configurable to your fairness definition and policy.
API
Legacy Integration
REST API compatible with EZD, SAP, and existing case management systems.
ON-PREM
Data Sovereignty
Can be deployed on your infrastructure – VPC, on-premise, or air-gapped. Designed for zero data egress.
People & Practice
The team behind explainable reasoning systems for regulated environments.
AdminLab.ai brings together administrative science, legal reasoning and AI engineering to deliver explainable decision flows for HR, policy, compliance, legal and risk teams.
// FOUNDING STORY
AdminLab.ai was founded to support traceable and fair decision-making in practice.
Initiated and funded by Mike. The goal: not another black-box model, but a structured reasoning engine that institutions and regulated firms can inspect, audit and improve.
Founding team
Yvonne
Administrative Science & Policy
Interface of administrative science, procedural fairness and policy design. Contributed work nominated for European Ombudsman Award 2025/2026.
AdminLab.ai Guru
AI Engineering & Reasoning
Designs structured reasoning architectures, implements logic flows for audit, reproducibility and institutional oversight.
Extended contributors & collaborators
LEGAL
Legal & Compliance
Case and policy analysis, mapping legal provisions into structured criteria, edge cases and exception handling.
RESEARCH
Research Partners
Academic and practice-based collaborators who stress-test reasoning chains and align models with emerging XAI standards.
OPS
Operational Support
Coordinates implementation with partners, manages feedback loops from prototype to operational use.
Explainability-first reasoning models
Treating explainability, reproducibility and fairness as primary design constraints, not afterthoughts.
01 · STRUCTURED
From narrative rules to explicit steps
Translate policies, legal texts and guidelines into explicit decision steps with clear conditions, thresholds and references.
02 · DETERMINISTIC
Same inputs, same outputs
Reasoning flows designed to be deterministic. Enables consistent application of rules and reliable re-analysis of past decisions.
03 · FULL LOGGING
Decisions come with a trace
Each decision logs which steps were evaluated, which conditions were met, and which policy references were applied.
04 · INSTITUTIONAL
Co-designed with practitioners
Models iterated with HR, compliance, legal and risk teams to reflect real workflows and the institutional understanding of fairness.
How the work has evolved
Pre-2024
Administrative practice & research: foundations in administrative science and legal reasoning.
2024
Founding of AdminLab.ai, initial framework development begins.
2025
European Ombudsman Award nomination, early models tested with practitioners.
2026 →
Operational deployments, governance frameworks, integration with institutional processes.
// GOVERNANCE & OVERSIGHT
"Independent challenge is a feature, not an obstacle."
AdminLab.ai is designed to operate under internal and external regulatory scrutiny. Logs reconstruct how a decision was reached; documentation ties rules back to originating policies.
internal auditlegal servicescomplianceexternal oversight
Designed for regulatory oversight & accountability
Combining administrative science, legal reasoning and AI engineering into a framework prioritizing explainability, auditability, and institutional trust.
European Ombudsman Award
Nominated for Good Administration Award 2025–2026. Grounded in established administrative standards and legal reasoning.
EU AI Act governance posture
Designed to support high-risk governance needs: documented rules, human oversight, auditability, and controlled deployment.
Resources & technical documentation
In‑depth materials for your due diligence and implementation planning.
Governance Infrastructure
Deterministic rule engine, audit trails, human‑in‑the‑loop – full architecture overview.
Request accessExplainable Decision Support
Use cases, methodology, and institutional benefits – concise overview for stakeholders.
Request accessImplementation Checklist
Step‑by‑step guide for procurement, integration, and governance readiness.
Request accessProcurement & due diligence pack
Download key documents for your procurement, legal, and security teams.
Executive brief
Request via email →
Technical architecture
Request via email →
Governance & oversight
Request via email →
Ready to deploy within your organisation?
We welcome conversations with HR, policy, legal, compliance and risk teams. Let's discuss your context and requirements.
Thank you! We will contact you within 2 business days.
Privacy Notice
We process your personal data only to respond to your inquiries and provide you with information about our services. We do not sell or share your data with third parties. For more details, please contact us at contact@adminlab.ai.
Terms of Use
This website is provided for informational purposes only. AdminLab.ai software is offered under separate enterprise agreements. By using this site you agree that no client relationship is formed. Jurisdiction: Warsaw, Poland.
GDPR Compliance
AdminLab.ai is designed to support compliance with the General Data Protection Regulation. In a typical deployment, we act as a processor and our customers remain controllers — the exact arrangement is defined in each enterprise agreement. Our DPO can be reached at dpo@adminlab.ai.