---
id: "action-track-provenance"
type: "action-item"
source_title: "Don't Let AI Slop Muck Up Your Company's Processes"
source_url: "https://hbr.org/2026/06/dont-let-ai-slop-muck-up-your-companys-processes"
source_timestamps: ["§ 1. Keep track of the provenance of unstructured data."]
tags: ["data-governance", "analytics"]
related: ["concept-unstructured-data-provenance"]
speakers: ["Matthias Holweg", "Thomas H. Davenport"]
action: "Document the origin of unstructured data to distinguish human ground truth from AI-generated content."
outcome: "Preserves authentic human signals for future analysis and prevents the blending of factual signal with AI noise."
sources: ["execution"]
sourceVaultSlug: "hbr-seg-execution"
originDay: 8
articleStem: "hbr-sig-54-ai-slop-processes"
sourceUrl: "https://hbr.org/2026/06/dont-let-ai-slop-muck-up-your-companys-processes"
sourceTitle: "Don’t Let AI Slop Muck Up Your Company’s Processes"
---
# Track Unstructured Data Provenance

**Action.** Implement systems to document the history of unstructured data, clearly distinguishing authentic 'ground truth' human information (e.g., raw interview transcripts) from AI-generated summaries or alterations. Record exactly what ground-truth data was used to generate any AI outputs.

**Outcome.** Preserves authentic human signals for future analysis and prevents the blending of factual signal with AI noise.

This is **Step 1** of [[framework-four-steps-knowledge-decay]] and the operational form of [[concept-unstructured-data-provenance]]; it presumes fluency in [[prereq-structured-vs-unstructured-data]]. Enforcement depends on detection capabilities the authors admit are weak — see [[question-detecting-ai-content]]. The enrichment overlay aligns it with NIST's guidance on tracking training-data provenance and metadata.


## Related across articles
- [[concept-unstructured-data-utilization]]
- [[action-deploy-genai-unstructured-data]]