Manual metadata tagging slowed teams down
Mad Engine employed a team of librarians whose full-time job was to tag incoming art assets. But even with that team in place, there wasn't enough capacity to tag every field. A t-shirt design might be tagged with "Captain America" but miss all the other characters on it. Those gaps made it harder for teams to find the art they needed and slowed the path from creative to sales.
Simple AI tools weren't enough
As AI exploded onto the market in 2023, Mad Engine began exploring which AI tools could help. What they found was a fragmented landscape. ChatGPT was strong on text. Gemini focused on images. Claude leaned toward code. But no single tool could handle the full range of their needs, and building separate AI solutions for each workflow was expensive and unsustainable.
Ad hoc AI wasn't the right strategy
"The initial approach was ad hoc," said Gary Gaffney. "We looked at each opportunity and tried to solve it individually. We recognized this is not sustainable long term. It was going to be very costly."
Mad Engine needed a platform, not just a tool. They needed something agnostic, centralized, and built to grow with them.
Protecting IP was non-negotiable
As a licensed apparel company, Mad Engine manages the intellectual property of major brand partners, as well as their own. Any AI platform they chose had to meet strict standards for security and data control. The AI had to work within their tenant, not share data outside of it.
One platform for every AI opportunity
Unlike point solutions tied to a single model, Vertesia supports all of the AI models. That means Mad Engine can choose the best model for each task, whether it's processing images, generating text descriptions, or analyzing production data. Vertesia acts as the hub. Everything else plugs in.
AI that works within existing workflows
Vertesia integrates directly into Mad Engine's existing app stack. AI was added on top of processes that already worked, making adoption faster and results more immediate.
Enterprise-grade security and governance
Vertesia operates as a secure, tenant-isolated SaaS platform. Mad Engine's data stays in Mad Engine's environment and it is never used to train any models. This gave leadership the confidence to move forward, knowing they could protect IP while still unlocking the power of AI.
A multi-phased approach designed to scale
Together, our teams designed a multi-phase AI deployment roadmap. Phase one focused on the highest-value, lowest-effort opportunity: automated asset ingestion. With phase one now complete, confidence to continue scaling is high.
Asset ingestion is now fully automated
Before Vertesia, a team of librarians manually tagged every incoming asset. Today, that process is 100% automated. An artist saves a new design. It flows into the digital asset management system automatically. It gets tagged. It gets vectorized. And it gets distributed to the rest of the organization with minimal human intervention.
Natural language search is a reality
Because every asset is now fully tagged and vectorized, teams can search for art the way they think. Instead of filtering through metadata, they can type: "show me everything with red polka dots." The system returns results. They can even upload an image and search for visually similar designs. This wasn't possible before.
"It's giving us a level of intelligence that we could never achieve manually," said Gaffney.
AI adoption is growing across the company
Early fears that AI would replace jobs have faded. Teams that once worried are now using the platform regularly. The reason? They saw it working. They saw it making their jobs easier, not threatening them.
"It's just another tool in your tool belt," Gaffney told his team. "It can really help you do your job better, maybe quicker or more successfully."
Production workflows have been automated downstream
Beyond asset tagging, Vertesia now helps Mad Engine automatically extract Pantone color data during ingestion. That data flows directly to import vendors, cutting out manual steps that used to slow production timelines.
Licensing approvals are moving faster
In phase two, Mad Engine is combining data from phase one to accelerate the art approval process with licensors. Assets go to partners more quickly and more accurately.
Speed to market has improved
Art gets tagged faster. It gets released to sales teams faster. That means it gets in front of buyers faster. Fewer delays. More deals.
"If a company wants to implement AI with trust and in a practical way, they really should work directly with Vertesia. They can help navigate the process, make the right choices, and bring it to fruition."
Gary Gaffney
EVP of Technology and Systems at Mad Engine