Does Validation Actually Help?

Two design critics watch the pipeline

Amplifier · Experimental · v8.1.0
The Frame

The doc-gen recipe validates every document — but nobody checked whether passing makes it better

Amplifier's document-generation recipe puts every generated doc through an automated validation pipeline from v0 to v3. It confirms the checks pass. It never confirmed that a design expert would actually call the result better.

First, is that pipeline even real? Look at what it actually runs.

A layered v0→v3 pipeline, inside a 3021-line recipe

The recipe's Version Flow runs structural validation first (ensuring sections exist), then parallel content checks, then parallel quality checks — substantial, genuine machinery.

It clearly does a lot. But does doing a lot mean doing good?

v0Initial generationraw output, before validation
v1Structural validationensures sections exist
v2Content checks (parallel)accuracy, completeness, instructions
v3Quality checks (parallel)depth, coherence, crossrefs, consistency, tone
The Tension

Passing the checks isn't the same as being judged good

Automated checks can confirm sections exist and rules pass. They cannot tell you whether validation quietly cost you clarity while it gained you structure. Validation may not be purely additive — and nothing in the pipeline was watching for that.

To see that, you need a different kind of reviewer.

Design Intelligence brings two expert agents

From Microsoft's design-intelligence bundle come two specialists: a voice-strategist and a layout-architect. Each brings a design expert's judgment the automated checks lack.

The trick is how they were wired in — without touching the document.

voice-strategist
Judges how the writing reads — voice consistency, clarity, and readability.
layout-architect
Judges how the document is structured — information architecture, content flow, and navigation.
Mechanism · Observe, Don't Touch

In v8.1.0, both agents were wired in as pure observers

The recipe is explicit: the feedback is purely observational — the agents make NO changes to the document. Their prompts say “Do NOT rewrite or fix anything.” Observation, not intervention, is what keeps the before/after honest.

So what exactly do they measure, and when?

The same named dimensions, scored 1–10, at v0 and v3

Each agent scores the raw v0 output before validation and the final v3 output after it — on named dimensions, persisted to di_v0_feedback.json and di_feedback.json.

Same scale, before and after — which sets up the comparison that matters.

voice-strategist (1–10)
voice consistency
clarity
readability
layout-architect (1–10)
information architecture
content flow
navigation

v0 vs v3 asks the question that turns faith into measurement

Report section 6.3, “v0 → v3 Comparison (CRITICAL FOR EVALUATION),” shows before/after scores per dimension — so the pipeline can finally reveal that validation is non-additive.

That third question is the whole point.

Which issues identified in v0 were addressed by validation?
Which issues persisted through validation?
Did validation introduce any new concerns?

Instrument your quality gates before you trust them

Shipped EXPERIMENTAL by Mollie Munoz on 2026-01-30 (PR #202), the lesson generalizes: treat your own validation as a hypothesis, not a guarantee. Measure whether it helps — don't assume it.

Faith becomes a measurement you can check.

v8.1.0
Recipe version (2026-01-30)
#202
PR by Mollie Munoz
+503
Lines added, 1 file changed
2
Stages observed: v0 & v3
Sources

Research Methodology

Data as of: local microsoft/amplifier checkout, HEAD dated 2026-03-02 (commit 676d677).

Feature status: EXPERIMENTAL — introduced in recipe v8.1.0 (2026-01-30).

Primary source: recipes/document-generation.yaml in microsoft/amplifier (3021 lines).

Research performed:

Gaps: The specific directional finding (that validation lifted navigation while costing clarity/readability) could NOT be verified — no di_feedback.json or di_v0_feedback.json run artifacts were found. The recipe scores those named dimensions and section 6.3 explicitly asks whether validation introduced new concerns, so the mechanism is verified, but observed values are not present in any local artifact.

Contributors: Mollie Munoz <mollie.munoz@microsoft.com> (authored PR #202 / commit 375aaed), co-authored-by Amplifier.

More Amplifier Stories