DAIMO

DAIMO is the narrow product. It answers one question: what data inside the institution is safe to expose to an AI system, and under what controls?

What it is

DAIMO catalogues the institution’s unstructured and semi-structured data — file shares, document stores, knowledge bases, ticket systems — and classifies each item against the institution’s own policy: PII present or not, sensitivity tier, regulatory category, data-residency obligations, retention rules. The output is a continuously updated catalogue plus an enforcement layer that AI systems (Copilot, internal RAG, partner LLMs) consult before retrieving a document.

The platform never moves data outside the perimeter. Classification runs inside the customer’s environment, against models the customer controls.

What ships today

DAIMO is in design-partner pilots with banks and a public-sector engineering team. Generally available is targeted for later this year; the design-partner status is documented honestly in every engagement.

The components shipping in pilots:

• Connector library for the document and ticketing systems institutions actually run
• Classifier pipeline with starter taxonomies (PII, PCI, regulated content)
• Policy DSL — rules expressed in plain config, version-controlled
• Lineage tracking — what classified what, with which model, against which policy
• Enforcement layer with adapters for Microsoft 365 Copilot, internal RAG endpoints, and partner LLM gateways

When to choose DAIMO

When the institution already has — or is about to acquire — a third-party AI system, and the procurement gate is “prove that the AI cannot retrieve what it should not see.” That is the pattern DAIMO is built for.

When the institution wants to build operational decisions on top of the same data, the answer is AlpOS, of which DAIMO-style classification is one layer.

How DAIMO and AlpOS relate

DAIMO is AlpOS’s classification layer, packaged as a standalone product for institutions that need only that layer. A team running AlpOS gets DAIMO’s classification as part of the platform — there is no second integration to do. A team running DAIMO standalone can migrate to AlpOS later without re-classifying.

What DAIMO is not

It is not a DLP product. DLP enforces at the network edge, looking at outbound traffic. DAIMO enforces at the data layer, looking at what an AI system is allowed to retrieve in the first place. The two are complementary, not substitutes.

It is also not a vendor-managed service. The classifier models, the policy rules, and the audit log all live inside the institution’s perimeter.