---
id: ins_andrew-jones-data-trustworthiness-over-output-metrics
title: 'AI products are completely limited by the data going in; output metrics mean nothing without trustworthy input data.'
operator: Andrew Jones
operator_role: 'Let me know if you agree or disagree:'
source_url: https://www.linkedin.com/feed/update/urn:li:activity:7372152232496041984/
source_type: thread
source_title: 'AI products are completely limited by the data going in; output metrics mean not'
source_date: 2026-04-10
captured_date: 2026-05-03
domain: [pmm]
lifecycle: []
maturity: applied
artifact_class: framework
score: { originality: 3, specificity: 3, evidence: 2, transferability: 3, source: 3 }
tier: B
related: []
raw_ref: raw/linkedin/reactions/linkedin-reactions-2026-04-10.md
---

# AI products are completely limited by the data going in; output metrics mean nothing without trustworthy input data.

## Claim
We're all obsessed with measuring the outputs of AI (accuracy, bias, drift etc) but if the data going in isn't trustworthy (or observable) then those outputs mean zilch, nothing at all.

## Mechanism
Data confidence is the missing link in AI observability. Making data observable and governable at the source ensures that downstream AI outputs are meaningful and reliable.