Work Relay

From campaign brief to production-ready assets.

Project Relay — A creative production workflow for multi-market campaigns
Category B2B Product Design / Creative Operations / AI Workflows
Status Self-initiated concept
Year 2026

Summary

  • Product strategy
  • UX architecture
  • Interface design
  • Design system
  • Data model
  • AI-assisted workflow

Relay is a concept B2B product for marketing and creative production teams who coordinate many campaign assets, variations, specifications and approvals across channels and markets. It replaces fragmented spreadsheets, folders and email threads with one production workflow: intake, requirements, production, review and handoff. Human-controlled, with AI assistance where it is genuinely useful.

Business context

A single campaign for a mid-sized e-commerce brand can require dozens of assets: social images, display banners, story placements, video thumbnails and landing-page creatives, each in several aspect ratios, languages and markets. The work is coordinated across design tools, spreadsheets, shared drives and email, and the state of any one asset is rarely clear.

Relay targets mid-sized marketing teams, creative production companies, performance agencies and multi-market e-commerce businesses. It deliberately avoids becoming a generic project manager or an all-purpose AI assistant.

The challenge

Illustrative scenario: a marketing operations team must deliver the “Autumn Collection Launch” in three markets (NL, DE, FR), two languages, across paid social, display and email, in multiple placements and aspect ratios, with several creative variants per placement. Specifications, ownership, versions and approvals have to stay coherent as the matrix grows.

The product has to make that matrix legible, keep every status trustworthy and support approvals that a brand or legal reviewer can rely on.

Product direction

Relay models a campaign as a structured set of requirements rather than a list of tasks. A brief is turned into a placement-and-dimension matrix; every cell becomes a trackable asset with a market, language, format, owner, due date, version and approval state.

The interface is calm and data-dense. A light surface, one deep green for primary actions, a lime signal for “ready” states and a small, consistent status vocabulary. Type is set in a single sans with tabular numerals so tables stay aligned.

AI assistance is scoped to preparation: parsing briefs into requirement drafts, suggesting format matrices, flagging missing metadata and summarising review threads. Every suggestion is visible, editable and confirmed by a person. Nothing publishes automatically.

Selected design details

Campaign intake

Name, channels, markets, languages, formats, source materials, deadlines and notes captured once.

Requirements matrix

The brief becomes a placement × market × language grid with specifications and naming conventions.

Production board

Every variant with status, owner, due date, version and approval state in one view.

Review and approvals

Preview, version history, comments and explicit approve or request-changes controls.

Export checklist

Approved assets, file specifications, naming validation and a package overview before handoff.

Status system

Five states, one colour each, used identically in tables, cards and notifications.

Systems and technical exploration

The demonstration dataset is internally consistent: one campaign, three markets, two languages, five asset types and a fixed set of placements. The same entities appear on every screen.

A credible MVP path: a React and TypeScript frontend, a relational data model (campaign, brief, placement, asset, revision, approval, comment), authentication with role-based permissions, object storage for files, an audit trail and asynchronous jobs for supported automation. AI integration is optional and controlled per workspace.

  • 01 React + TypeScript frontend
  • 02 Relational data model
  • 03 Role-based permissions
  • 04 Object storage + audit trail
  • 05 Async processing for automation
  • 06 Optional, user-controlled AI assistance
  • 07 Data retention and access policies

Next-stage opportunities

  1. 01 Build a self-contained interactive demo of the production board and review flow with mock data.
  2. 02 Validate the requirements-matrix concept with a small number of creative operations practitioners.
  3. 03 Prototype the brief parser as an assistive draft with explicit confirmation controls.

Concept project. Independently initiated. Not a client commission. All names, data and content are fictional.

Next project

Seasoned

Brand Identity / Packaging / E-commerce

Related service

Creative Technology & AI

Explore practical ways to connect interfaces, automation and emerging technology.

Have something worth building?

Let’s shape it into something distinctive.