Brand asset production at scale runs on two systems. The brand system itself, the master logos, type tokens, color tokens, and component libraries, stays in Adobe Creative Cloud and Figma. The production queue, the row-per-asset record of who requested what, which agent generated the variation, which brand lead checked it against the system, and which reviewer approved the final, lives in Dock. The Dock row is the durable artifact. The platform file is the canvas the row points at.
Figma, Adobe Creative Cloud, Sketch, Webflow, and Framer remain the system of record for the canvas: frames, layers, components, the source files. Dock is the system of record for what the team and the agent interpret around the canvas: the brief, the variation rationale, the brand-system check, the reviewer decision, the production-asset queue. Each Dock row carries a pointer to the platform record (figma_file_key, figma_node_id, aem_asset_id, webflow_collection_id), the agent identity, the reviewer, and the timestamps that bind them. The agent re-fetches platform data on each pass to read current state. Dock holds the persistent interpretive layer that survives sessions, handoffs, and quarter boundaries. This is the same architectural split documented in the broader Dock for Design overview and in the deeper Dock + Adobe Creative Cloud workflow.
The asset-production queue
| asset_id | request | platform_pointer | variations | brand_check | reviewer_decision | status |
|---|---|---|---|---|---|---|
| ASSET-2041 | Q3 paid social, 5 sizes, "Series B announcement" | figma_file: aB3..., node: 12:441 |
4 generated by agent:ved |
passed, tokens match, type scale ok | approved by lila@, 2026-05-28 14:02 |
shipped to AEM |
| ASSET-2042 | Customer story, hero + 2 supporting | figma_file: aB3..., node: 12:512 |
3 generated by agent:ved |
failed, headline color off-token | returned for rework | in revision |
| ASSET-2043 | Event banner, 3 placements | aem_asset_id: urn:aem:b91... |
2 generated by agent:ved |
passed | approved by marc@, 2026-05-29 09:10 |
scheduled |
The columns are load-bearing. platform_pointer is what the agent re-reads to confirm the current state of the source file. brand_check is the brand lead's structured judgment, not a vibe. reviewer_decision carries the human who took accountability, which is the part audits and post-mortems care about.
One asset, end to end
Marketing files ASSET-2041 in Dock: Series B announcement, five paid-social sizes, deadline Thursday. The agent reads the brief, fetches the brand library from Figma via API, generates four variations across the requested aspect ratios, and writes the candidate frames back as a new Figma page with a link in the Dock row. The brand lead opens the row, runs the brand-system check (token match, type scale, logo clear space) and marks brand_check: passed. The reviewer opens the row, approves variation 2, and the agent pushes the approved asset to Adobe Experience Manager with the Dock asset_id written into the asset metadata. The full chain from brief to ship is one row, each step attributed, each handoff dated. The mechanics of the brief and review step are covered in design brief and review.
Why this matters
Attribution survives the asset. Six months after shipping, someone asks who approved the off-system headline color on the Q3 campaign. The Dock row answers in seconds. The Figma file alone cannot, because Figma records the last editor, not the chain of decisions that produced the final. This is the same logging substrate described in agent audit and compliance.
Handoff stops costing a week. When the brand lead rotates off the team, the new lead reads the last forty rows and learns the actual standard the team applies, not the one written down two years ago. The queue is the institutional memory.
Design teams get a daily driver that does not fight their tools. The canvas stays in Figma and AEM, where the muscle memory lives. The queue, the conversation, and the agent live in Dock, where the durable record belongs. The agent works under its own agent identity, and any push to production is gated by the dangerous ops contract so an unapproved variation cannot ship by accident.
Start with one asset queue, one brand lead, one reviewer: open a Dock workspace.
FAQ
Q: Does Dock replace our DAM? A: No. Adobe Experience Manager Assets, Bynder, or whichever DAM you run stays the source of truth for the finished, approved file. Dock is the production queue that produces the asset and writes the decision trail. The Wikipedia DAM overview covers the core DAM functions Dock does not duplicate.
Q: How does the agent stay on-brand?
A: It reads the brand library directly from Figma or Adobe Creative Cloud on each pass, using the documented Figma components, styles, and shared libraries model. The brand lead's brand_check row is the human gate that catches what the API read missed.
Q: What about brand consistency across markets and agents? A: Consistency is a function of the queue, not the canvas. Every variation flows through the same brand check and reviewer row, so the standard applied in March is the standard applied in November. The classical brand management literature frames this as controlling perception across all touchpoints, which is exactly what a row-per-asset queue enforces.
Q: Can multiple agents work the same queue?
A: Yes. Each agent writes under its own identity, so the asset row shows agent:ved generated variations, agent:mei wrote copy, lila@ approved. The reviewer always sees who did what.