Data and intelligence

Scope

This page owns how real work becomes governed intelligence: operational traces, evaluation artifacts, review trails, telemetry, rights, privacy, aggregation, and brokerage. It does not grant permission merely because data passes through the runtime.

Current answer

Zenith sits at the orchestration layer where requests, context, tool calls, intermediate decisions, reviews, corrections, and outcomes can be observed. That position can create high-signal domain data, but only when the architecture separates private operations from licensable derivatives and performance telemetry.

The required data unit is not raw input/output. It is:

task context + output + rubric version + reviewer decision + error taxonomy + correction + final outcome + rights envelope

Every artifact must carry ownership, approved purpose, lineage, retention, revocation, redistribution, attribution, and revenue-routing metadata. Without this, Zenith is a centralized capture layer with sovereignty branding.

Evidence and influences

• defines the capture position: Zenith's orchestration layer is the data capture position — every harness IO passing through it becomes anonymized training data for specialized LLMs in prompts represent a context defining thought.
• qualifies the capture claim: micro1 business model and data layer — full report 2026-08-28 - raw workflow traces require specification, transformation, review, telemetry, rights, and delivery.
• defines the economic model: Zenith's data brokerage model captures 10 percent of a specialized operational data market that doesn't yet exist as a product.
• defines the sovereignty distinction: micro1 clarifies Zenith's differentiator as a self-sovereign brokered data mesh rather than a centralized data vendor.
• defines the market category: Operational workflow exhaust is becoming licensable AI training data.

Three data planes

1. Private operating data: required to run the organization's workflows; not automatically available for training or sale.
2. Derived training and evaluation artifacts: rights-cleared outputs that may be used for improvement, research, or licensing.
3. Performance telemetry: quality, reliability, routing, cost, drift, and reviewer data used to improve the system.

Permission does not silently flow from one plane to another.

Acquisition channels

Zenith's structural advantage is continuous operational capture at the source. A complete foundry also needs explicit modes for:

• bounded historical corpus contribution;
• continuous workflow contribution;
• expert-authored prompts, rubrics, references, and evaluations;
• contextual production-agent evaluation;
• machine-generated enrichment with human verification;
• partner-provided physical or robotics data where strategically relevant.

micro1's current coverage is documented in micro1 — full extraction report 2026-08-28 and micro1 business model and data layer — full report 2026-08-28.

Rights envelope

Every governed artifact should carry:

• source owner and contributor attribution;
• consent scope and approved buyer classes;
• raw, derived, synthetic, or aggregated status;
• transformations and model versions;
• rubric, review, correction, and lineage history;
• retention, deletion, and revocation behavior;
• redistribution and multi-license eligibility;
• compensation and royalty routing;
• jurisdiction, sensitivity, and re-identification risk.

Quality and evaluation loop

• procedural agent harnesses with an overseer layer let process authors configure human review points within Frank processes - review belongs in process definitions.
• Frank is currently a task agent without a meta-agent layer — the hyperagent architecture identifies this as the bottleneck that prevents compounding self-improvement - telemetry alone is insufficient without causal improvement.
• micro1's quality layer routes work through peer review, operations review, client review, and risk-triggered rework.
• micro1's workflow telemetry is a second-order dataset that predicts who can produce reliable expert data.

Privacy and compliance

• GDPR compliance in Zenith — report and implementation requirements.
• the hub intake boundary is the natural GDPR compliance gate — personal data is classified consented encrypted and routed at a single point.
• personal data must never enter git — only anonymized identifiers belong in version control so right to erasure is a clean database delete.
• self-sovereign personal data encrypted with the user's public key means the platform operator cannot be compelled to produce readable data.
• GDPR applies based on the residency of data subjects not the location of the server.
• Canada's PIPEDA adequacy decision with the EU gives Canadian-operated services a clean legal path for processing EU personal data.
• psychographic profiling is special-category data under GDPR Article 9 requiring explicit documented consent.

Anonymization is not treated as a single sufficient control. Rare workflows, terminology, timestamps, and decision patterns can remain identifying after direct PII removal.

Model and network implications

• Zenith's competitive moat against AI monopoly is owning the orchestration layer that generates training data for open-source foundation models.
• Communities monetize their knowledge bases by sharing them with AI agents as a fee-based API.
• specialized cybernetic business infrastructure can eventually be compressed onto a chip making an entire business representable as specialized silicon.

The model-training and silicon arc is long-term. The immediate proof is a rights-cleared evaluation and training artifact generated from a live organizational workflow.

Open questions and tensions

• What exact storage boundary holds rights metadata and derivative lineage across Git, databases, and object storage?
• How does revocation propagate to already transformed, aggregated, or licensed artifacts?
• What review depth is required by risk class and buyer use?
• How are contributor economics calculated when an artifact has many upstream sources?
• Which first data product has a named buyer and a sufficiently dense source community?
• What independent audit proves privacy, quality, and chain of title?

Related canonical pages

• Zenith Research - strategy and business model - brokerage and revenue layers.
• Zenith Research - system architecture - capture and persistence implementation.
• Zenith Research - market and competition - micro1 and market validation.
• Zenith Research - network governance and economics - source participation and settlement.