Listing Platform Product Review Sandbox

Atlas HydrationUnited States

Deep implementationDIR-014

Sandbox v1.3 · Stateful vertical slice

Move concentration proof into Image 2H2Go Portable Hydrogen Water Bottle · B0H2GO1001

Current: Recommendation

1Recommendation 2Roadmap 3Execution 4Workspace 5Review 6Approval 7Publishing 8Verification 9Measurement 10Learning

Recommendation decision

A recommendation is explainable evidence, not an instruction

AI recommendationHigh impact87% confidence

Move hydrogen concentration proof into image 2

Top converting competitors establish measured concentration before lifestyle benefits.

ScopeProduct · US marketplaceExpected resultImprove generic-search conversionEstimated effort1 strategy + 1 image revision

Decision stateNew

Evidence used in the decision

Each claim remains traceable to its source

SearchGeneric traffic is growing

“Molecular hydrogen water bottle” grew 21% and converts above the broader portable-device cluster.

CompetitiveProof appears before lifestyle

Seven of the ten highest-converting competitors establish concentration or test proof in Images 2–3.

VOCShoppers question whether the device works

Review language repeatedly asks how concentration is measured and whether results vary by water source.

ComplianceProduct-performance claim is supportable

The evidence packet supports a concentration statement when the test method and conditions remain visible.

What the system will create

Roadmap itemMove concentration proof into Image 2Owner: Creative Strategy · Priority: High

→

ExecutionGraphic Strategy RevisionInputs frozen when launched

→

OutputApproved Graphic Strategy versionPublishing remains separately gated

What this page is supposed to answer

What should we do next across the catalog?
Which work is already complete, running, blocked, or recommended?
Which execution should be added, reordered, deferred, or launched?
What should a Listing Director, Account Manager, Brand Director, Creative Director decide or understand here?

Why it exists

Director is the command center: it decides what should happen, in what order, at what scope, and whether work should be manual, assisted, or machine-run. This specific page exists to review explainable AI recommendations, evidence, impact, readiness, priority, and affected scope before adding work to the roadmap.

How to use it

Start with the page header to confirm the screen, scope, and review status.
Read the primary cards or table as the operating surface for the current decision.
Use the element info buttons to inspect why a region, component, action, or state exists.
If the page feels wrong, add a Review Mode call-out and then save an explanation update describing what changed and why.

Important nuance

The roadmap is the center of Director because AI recommendations should be advisory, while the listing director controls sequence and priority.
The catalog grid has three lenses over the same products: current assets, audit and quality, and execution readiness.
AI guidance belongs where action decisions happen, not as a disconnected chatbot.
Used when the core job is finding, filtering, comparing, and acting on many records.
Old and new explanations should both remain available because product decisions are part of the system history.

Current explanation version

Architecture baseline2026-07-16 / Architecture conversation

What: Initial product guidance for AI Recommendations.

Why: Generated from the Listing Division architecture, screen map, navigation spec, execution ontology, component system, and sandbox implementation.

Guidance: AI Recommendations should review explainable AI recommendations, evidence, impact, readiness, priority, and affected scope before adding work to the roadmap.

recommendation cardswhy nowevidencepriority scoreimpact estimateaccept/modify/defer/dismiss

Architecture signals

Screen map: DIR-014Surface: DirectorTemplate: CollectionPageTemplateEntities: roadmap, execution, recommendation, review_request, impact_analysisStates: New, Accepted, Modified, Deferred, Dismissed