---
title: "Data Strategy: K-Framework Layer A.02"
description: "Validate the Data Strategy layer of an AI system: contract-first data, lineage from source to retrieval, schema-level PII gating, and reviewed retention. A CEO/CTO field guide."
source: "https://www.kensink.com/k-framework/data-strategy/"
canonical: "https://www.kensink.com/k-framework/data-strategy/"
---
[‹](/k-framework/system-design/ "System Design") [›](/k-framework/algorithmic-fundamentals/ "Algorithmic Fundamentals")

[← The K-Framework](https://www.kensink.com/k-framework/)

★ Foundations · Systems Thinking Layer 2 of 16 Visual guide

PILLAR A · LAYER 02 · A.02

# Data Strategy.

Collect, structure, govern.

What a CEO/CTO needs to know  
The fastest model on earth cannot recover from bad data discipline. Ask to see the data contract before you ask to see the demo.

Source to contract to retrieval, with PII tagged at the schema boundary so it never reaches the prompt.

\[WHAT IT IS\]

## The engineer’s view, in plain language.

Data is the substrate of every AI feature. Schema, lineage, retention, and consent boundaries get defined before we train, fine-tune, or prompt. Discipline here is what keeps PII out of a completion and lawyers off the phone.

\[HOW WE BUILD IT\]

## What “done right” looks like.

01

### Data contract first

Schema, ownership, and allowed uses are written down before ingestion, so the system never quietly depends on a field nobody agreed to.

02

### Lineage end to end

Every value is traceable from source through retrieval, so when an answer is wrong you can find which input made it wrong.

03

### PII tagged and gated

Sensitive fields are tagged at the schema level and blocked at prompt assembly. Retention matches the firm's compliance window, reviewed, not assumed.

\[ MATURITY LADDER \]

## Where does your build sit?

Four rungs from absent to production-grade. Level 3 is the target, and the only one that survives a real production incident.

L0

Absent

Data is ingested as found. No schema, no lineage, no PII boundary.

L1

Ad-hoc

Schemas exist per source but drift, and PII handling is manual and inconsistent.

L2

Managed

A data contract and retention policy exist, but lineage and PII gating are partial.

L3 Target

Production-grade

Contract-first data, lineage from source to retrieval, PII tagged and gated, retention reviewed against compliance.

\[VALIDATE IT YOURSELF\]

## How to check it’s really there.

You do not need to read the code. Ask these questions and demand these artifacts. Vague answers are the finding.

★ Ask your team

-   ? Where is our data contract, and what does it forbid?
-   ? If a completion is wrong, can we trace which source field caused it?
-   ? What stops PII from reaching the prompt, and who reviewed the retention window?

★ Demand to see

-   A written data contract with schema and allowed uses
-   A lineage map from source through retrieval
-   Schema-level PII tags + a prompt-assembly gate, plus a reviewed retention policy

● WHAT L0 LOOKS LIKE

## The failure mode, in production.

Ingesting whatever is available and hoping the model cleans it up. RAG over a hairball. PII surfacing in completions. Lawyers calling.

Useful for a CEO or CTO sizing up an AI build? Share the Data Strategy layer.

Share[](https://twitter.com/intent/tweet?url=https%3A%2F%2Fwww.kensink.com%2Fk-framework%2Fdata-strategy%2F&text=Data%20Strategy%3A%20K-Framework%20A.02%20%C2%B7%20Kensink%20Labs)[](https://www.linkedin.com/sharing/share-offsite/?url=https%3A%2F%2Fwww.kensink.com%2Fk-framework%2Fdata-strategy%2F)

[View .md](https://www.kensink.com/k-framework/data-strategy.md)

[← Layer 1 System Design](https://www.kensink.com/k-framework/system-design/) [Layer 3 → Algorithmic Fundamentals](https://www.kensink.com/k-framework/algorithmic-fundamentals/)

## Want this layer audited in your stack?

We run the K-Framework against your AI build and hand you the gap list, ranked by what it will cost you in production.

[Book a K-Framework audit →](https://www.kensink.com/contact) [All 16 layers](https://www.kensink.com/k-framework/)