---
schema_version: "newruntime-agent-readable-v0.1"
type: "raw_signal"
id: "tg-2690"
slug: "ai-feedback-needs-behavioral-traces"
title: "AI product feedback needs behavioral traces"
description: "For nondeterministic products, edits, retries, overrides, abandonment, and sampled outputs reveal more than isolated ratings."
observed_at: "2026-07-14"
why_it_matters: "Teams need to preserve the prompt, context, output, and next user action so feedback can become an eval case instead of an anecdote."
novelty: "notable"
verification_level: "source-inspected"
signal_type: "field-report"
evidence_kind: "mixed"
status: "published"
telegram_message_id: 2690
telegram_url: "https://t.me/qwgai/2690"
topics: ["product-feedback","evals","human-traces"]
entities: ["Jeff Gothelf"]
related_patterns: []
source_urls: ["https://jeffgothelf.com/blog/how-to-collect-product-feedback-when-your-ai-gives-every-user-a-different-answer"]
import_batch: "telegram-2026-07-17-v1"
routes: {"html":"https://newruntime.com/signals/ai-feedback-needs-behavioral-traces/","markdown":"https://newruntime.com/signals/ai-feedback-needs-behavioral-traces.md","json":"https://newruntime.com/signals/ai-feedback-needs-behavioral-traces.json"}
source_format: "telegram-export-normalized-json"
---

# AI product feedback needs behavioral traces

## Observation

For nondeterministic products, edits, retries, overrides, abandonment, and sampled outputs reveal more than isolated ratings.

## Why it matters

Teams need to preserve the prompt, context, output, and next user action so feedback can become an eval case instead of an anecdote.

## Provenance

Normalized from QWG AI Telegram message 2690. The original Russian-language record remains available at https://t.me/qwgai/2690.