Why Sentra Exists
We are building the infrastructure layer for AI governance. This is what we observed, what we decided to build, and why we are starting in Europe.
What We Observed
EU B2B companies are deploying AI systems faster than they can govern the decisions those systems produce. Hiring tools, risk scoring models, contract analysis systems, customer-facing automation — these are in production, influencing real decisions, with no structured record of what they do or who is responsible for them.
The problem is not technical. Engineering teams can build and deploy AI systems. The problem is organizational: there is no structured layer for documenting, classifying, and governing the decisions those systems produce. When an enterprise buyer or regulator asks a basic question — "what decisions do your AI systems make and who oversees them?" — the answer is assembled from Slack threads, wikis, and memory.
This gap compounds with every system deployed. Inconsistent risk assessment. No structured record of decisions. Procurement questions that take weeks to answer. Sales cycles that stall at the security review stage.
Sentra is the infrastructure layer that closes this gap: a structured registry, a risk classification framework, decision-level oversight, and procurement-ready reports.
Why We Are Starting in Europe
The EU AI Act entered into force in August 2024. Key provisions apply progressively: February 2025 brings prohibited AI practices and AI literacy obligations; August 2025 introduces governance obligations for general-purpose AI models; August 2026 makes most AI Act requirements applicable; some transition cases extend into 2027. Sentra does not replace legal interpretation. Sentra helps companies structure the documentation needed to govern AI systems responsibly.
But the regulatory deadline is not the primary driver for most of our design partners. The primary driver is procurement. EU B2B companies selling to enterprise customers are already being asked to document their AI systems. These questions are not hypothetical — they are blocking deals today.
We are starting in Europe because the combination of regulatory pressure and enterprise procurement scrutiny creates the clearest, most immediate need. The governance infrastructure we are building here will be relevant globally — but Europe is where the problem is most acute right now.
We are not a compliance consultancy. We are building software infrastructure. The EU AI Act is context, not the product.
The Synthetic Workforce Problem
Companies already govern: people, money, infrastructure. But modern organizations also operate a synthetic workforce of AI systems that influence decisions. Sentra provides the governance layer for that workforce.
Companies now operate a synthetic workforce alongside their human one: AI systems that screen candidates, score credit risk, draft contracts, route support tickets, and flag fraud. These systems make or influence decisions at scale, continuously, without a manager reviewing each one.
Unlike employees, these systems have no org chart entry. No performance review. No documented scope of authority. When something goes wrong — or when a buyer asks who is responsible — the answer is assembled from Slack threads and memory.
This is not a monitoring problem. Monitoring tells you if a system is running. Governance tells you what the system does, who owns it, and how it influences decisions. These are fundamentally different questions.
Registry
A centralized inventory of every AI system in production — use case, model, owner, decision relevance.
Risk Classification
A consistent framework for assessing impact, automation level, and regulatory requirements.
Decision Logging
Structured record of AI-influenced decisions, queryable and exportable.
Governance Reports
Procurement-ready documentation generated from the registry, not assembled by hand.
How We Are Building This
Infrastructure, Not Consulting
We are building software, not a services business. The goal is a governance layer your team can operate independently: a registry you maintain, a classification framework you apply, decision logs your system generates, and reports you export on demand. We help you set it up. You run it.
Vendor-Neutral by Design
Sentra does not train models on your data. We do not compete with your AI systems. We sit above your existing infrastructure — OpenAI, Anthropic, AWS Bedrock, Mistral, or custom models — and provide the governance layer on top.
The Category We Are Building
AI governance infrastructure does not yet exist as a defined software category. Most companies are solving this problem with spreadsheets, internal wikis, and manual processes. We believe this will be a standard layer in the enterprise software stack within three years. We are building it now.
We Are Early-Stage
Sentra is in active development. We are working with a small number of design partners to build the product alongside real governance problems. We do not have a large customer base or years of production data. What we have is a clear thesis, a working product, and direct founder involvement in every implementation.
Why This Founder, Why Now
Kyrill has spent the last decade working directly with startups and scale-ups across Europe. This experience created a unique vantage point: observing how AI adoption accelerates faster than governance infrastructure, and how this gap becomes a blocker for enterprise deals.
He has seen the same problem repeatedly: companies deploying AI systems with no structured way to document them, classify them, or answer basic procurement questions. He has watched sales cycles stall at security reviews. He has observed how European companies face regulatory pressure earlier than the rest of the world.
This is not a theoretical problem. It is a daily friction point for the companies he works with. The founder-market fit is direct: deep access to the European startup ecosystem, direct observation of procurement friction, and the credibility to build infrastructure that founders and enterprises trust.
Europe is the right starting point because regulatory pressure (EU AI Act) and enterprise procurement scrutiny converge here first. The governance infrastructure built for European companies will be relevant globally — but Europe is where the problem is most acute and where the founder has the deepest relationships.