VML Bloodstream
01 / 14
Every idea has
a heartbeat.

Ideas deserve to live. They pulse through briefs and handoffs, through brainstorms and late nights, evolving as they go. Bloodstream exists to keep the pulse strong.

Exco Overview — March 2026
The challenge

What we're solving

Briefs arrive inconsistent and incomplete

No visibility into pipeline, quality, or resource pressure

Reverts and rework cost production time nobody tracks

The platform

The agency's
circulatory system

AI-powered brief management and delivery platform — from intake to production tracking. Built specifically for VML SA.

Discover
Define
Develop
Deliver
Diagnose

Integrated with Workbook • OKTA SSO • Box.com • Microsoft Graph • N8N

The journey

Built on proven foundations

Antonio's Briefing Tool

  • Functional brief upload prototype
  • AI enhancement workflows
  • Complexity framework (Green / Orange / Red)
  • Section-by-section AI prompts

Bloodstream Platform

  • Production-grade infrastructure
  • 10+ system integrations
  • Full brief-to-delivery lifecycle
  • Client Portal deployment
  • Quoting & scheduling tools
  • Governance & audit trails
What's shipped

Live on Production

Briefing
Client PortalSeparate client-facing deployment
AI Brief Enhancement3 complexity levels • section-by-section
Pulse CheckAI quality scoring before finalisation
Vision ExtractionImage-heavy & scanned brief support
Delivery & Production
Revert SystemAI adjudication • responsibility & severity tracking
Studio Portal (Axon)Internal brief tracker across teams
Platform & Governance
Circulation DashboardPipeline metrics • AI impact • leaderboards
Admin & PermissionsRole-based access control
OKTA SSO Authentication
Workbook Data SyncDaily automated client & employee sync
Box.com IntegrationAuto folder creation & brief upload
Bug Logging & User GuidesIn-app FAQs • onboarding • feedback system
In testing

Built & In Testing

Preview

Meeting Magician

AI meeting scheduler with Microsoft Graph calendar integration and Teams meeting creation.

Complete 100%
Preview

Resource Planner

Annual campaign calendar with heat-mapped resource pressure and overlap detection.

Complete 100%
In Testing

Internal Briefs

6 brief types for internal workflows — Production, Events, Radio, Stills, Renewal, Agency.

Complete 90%
In Testing

Quoting Tool

4-step wizard with AI scope matching for accurate project quoting and estimates.

Complete 85%
Preview

Timing Plan Generator

AI-generated Gantt charts from brief context. Drag-to-reschedule, conflict detection, PDF export.

Complete 75%
Underway

Scheduling

Resource allocation timeline. 443 resources, 33,000+ jobs, 3,800+ bookings synced from Workbook.

Complete 50%
What's next

The Roadmap

Now

Scheduling Write-back

Create, edit, and delete resource bookings directly to Workbook

Drag-and-drop Allocation

Visual resource management with real-time capacity heatmaps

Then

QA & DTP Ticketing

Structured ticketing process for QA reviews and DTP production

Then

Email Notifications

Automated notifications to teams based on their function and workflow stage

Later

Client Health & Reports

Vitals dashboard — burn rates, deadlines, timesheet compliance, scheduling reports

Deep dive

Scheduling

What we have today

  • Full Workbook API mapped and synced
  • All 443 employees with roles and departments
  • 33,000+ jobs with project data
  • 3,800+ existing bookings visible
  • Daily capacity and time-off data
  • Timeline view with filtering and search

What we're building now

  • Write-back to Workbook — schedule from Bloodstream
  • Availability checking per role (e.g. find available copywriters)
  • Sort and plan against real capacity
  • Single view of all agency resources and gaps
  • Clear visibility: who is busy, who is not
  • Traffic can plan properly using real data

The challenge: Workbook isn't widely used for scheduling yet — many teams use other tools, so current booking data lacks timed hours. This is a process change that comes with adoption.

Under the hood

Production-ready, not prototype

47 tables
Database architecture across 2 Supabase projects with row-level security
46 migrations
Versioned, sequential database migrations — structured and repeatable
100k+ records
Synced daily from Workbook — resources, jobs, bookings, capacity
Full audit
Every AI interaction logged — model, tokens, cost, execution time
Workbook API
Reverse-engineered undocumented API with 7+ Databoard integrations
OKTA SSO
WPP authentication with non-standard JWT handling and session management
By the numbers

Scale of the build

72,400+
Lines of Code
216
Components
47
Database Tables
31
Pages
10+
Integrations
345
Commits
43
Env Variables
3.5
Months
Effort comparison

What this would have taken

14+ months
Traditional development
1 senior full-stack developer
3.5 months
AI-augmented development
1 developer + Claude

~4x acceleration — production enterprise software, not throwaway code

Philosophy

AI isn't a feature.
It's the architecture.

Brief enhancement
Strategic advice
Quality pulse check
Timing plan generation
Revert adjudication
Meeting scheduling
Quote scope matching
Vision extraction

And the platform itself was built with AI.

Experience

See it in action

Every idea has a heartbeat.
Bloodstream keeps it beating.
VML Bloodstream