Case Study VEYLAN

An AI production engine for marketing, on one backend.

VEYLAN is an AI-native operating system for advertising. Teams set a Master Brief, run governed workflows across strategy, creative and media, review AI recommendations with evidence, and export the deliverables. We built the backend and API that hold all of it together.

Client
VEYLAN
Sector
Advertising · AI OS
Engagement
Backend & API platform
Stage
Live 2026
Platform
Web · API
The VEYLAN homepage hero reading An AI production engine for marketing execution
One workspace for advertising: a Master Brief, governed workflows, and deliverables that export to PPT, PDF, Excel and Word.
By the numbers
70%Cost reduction at PadSquad
10×Faster campaigns at PadSquad
20-30Point tools replaced
Multi-tenantAI backend

An ad team runs on 20 to 30 tools that never talk to each other.

Strategy, creative, media buying and reporting each live in their own tool. A strategist pulls numbers from one place, writes in another, hands off to designers in a third, and someone re-uploads the result to a DSP.

Every handoff costs time and attention. The coordination overhead stretches cycle time, and the margin a campaign should earn leaks out of the gaps between the tools. The work gets done, but the stack works against the team doing it.

VEYLAN set out to collapse that stack into one workspace. To make that real, the platform needed a backend that could model the whole business and run AI work without making people wait.

One backend that models the agency and runs the AI behind it.

Zain built the backend that powers the platform: a multi-tenant Flask REST API spanning organizations, agencies, clients, brands, campaigns, strategies, proposals, creative briefs and conversations. One data model carries an agency, the clients it serves, and every campaign in flight.

What that meant in practice

  • A multi-tenant Flask REST API across organizations, agencies, clients, brands, campaigns, strategies, proposals, creative briefs and conversations
  • Long-running AI work, including story writing, chatbot conversations and campaign creation, runs through asynchronous workers over RabbitMQ
  • Real-time progress streams back to the workspace over Pusher, so a running job shows its state instead of a spinner
  • Integrations into external AI services, AWS S3, advertising data warehouses and DSPs feed the workflows
The VEYLAN backend architecture diagram showing the Flask REST API, RabbitMQ async workers, Pusher real-time updates and integrations to AI services, AWS S3 and DSPs
Architecture

Slow AI jobs leave the request path.

A creative brief or a campaign build can run for minutes. We push that work onto asynchronous workers over RabbitMQ, so the API answers fast and the heavy step runs on its own. Pusher streams progress back as each stage finishes, and the workspace shows what is happening in real time.

Async workers · RabbitMQ + Pusher
See it work

Walkthrough: a Master Brief moving through governed workflows to exported deliverables

The VEYLAN campaign workflow diagram tracing a Master Brief through media planning, creative briefs, budget recommendations and performance reporting to exported deliverables
Workflow

A brief in, deliverables out, a person in the loop.

The Master Brief feeds governed workflows for media planning, creative briefs, budget recommendations and performance reporting. The AI proposes, attaches the evidence behind each recommendation, and waits. A person approves what moves forward, then exports the result to PPT, PDF, Excel or Word.

Governed workflows · evidence-backed
The hardest part

One data model has to keep an organization, its agencies, their clients and every brand and campaign apart while the same async pipeline writes AI output into all of them at once. We made tenancy a property of every record and every queued job, so a strategist on one account never sees another account's work in flight.

Ego Eimi Engineering note · the build

Live, with an early agency reporting real savings.

PadSquad, an early adopter agency, reports 70% cost reductions and campaigns running 10x faster on VEYLAN, work that used to take weeks now landing in minutes.

VEYLAN launched from stealth in June 2025 and debuted its full platform in October 2025 at Advertising Week NY. WARC named it a category leader in its MAMMA research, and the company acquired Advisr in September 2025. The backend we built runs underneath all of it.

Python Flask RabbitMQ Pusher AWS S3 PostgreSQL
Our role
Backend & API platform
Team
Lead + specialists
Model
Product build
Status
Live
Project partner Zain Raza
HYSS crew booking platform for AV and live events
AV & live events →

HYSS: crew booking without the spreadsheets.

Book, schedule and pay freelance AV crew in one platform, conflicts caught before they happen.

Read the case →
PLEXOR vision AI reading retail camera footage and POS data
Retail · Visual AI →

PLEXOR: existing cameras turned into a profit engine.

Vision AI that reads live footage and POS to catch theft, coach staff and flag hazards.

Read the case →
All case studies 33 builds

No tech team of your own? See how we work, compare in-house vs an agency, or read the build-without-a-tech-team guide.

Next / Your build

Taking on new builds

Have something in mind?

Tell us what you're making. We reply within a day with a fixed price and a date.

Start a conversation More case studies