As Plex expanded its global AVOD platform, it faced a critical challenge. How do you deliver relevant discovery, personalization, and monetization across tens of thousands of titles when the underlying content data is inconsistent and lacking the depth and quality of metadata required to power a modern streaming experience?
By partnering with Vionlabs, Plex built a scalable, AI-driven foundation for the generation of high-quality metadata at scale. This enabled a deeper understanding of content that now powers engagement, discovery, and revenue across its platform.
“Vionlabs helped us deliver a better user experience at scale for all Plex users around the world.”
The challenge was not just scale. It was understanding
When Plex launched its AVOD service globally, the ambition was clear: to build a platform where users can easily discover and enjoy content from a vast and diverse catalog.
But as the catalog expanded, a more fundamental problem emerged.
Content was arriving from multiple distributors, across regions, across decades, and across formats. Each asset came with metadata shaped by different standards, priorities, and levels of detail.
The result was not only incomplete data, but also metadata that lacked depth and true understanding of each asset. What existed was often inconsistent at a structural level, making it unreliable as a foundation for discovery and personalization.
At its core, the problem was not just about missing data. It was about missing understanding.
Without consistent and complete metadata, it becomes almost impossible to build reliable systems for recommendation and personalization. Building a recommendation engine on top of incomplete metadata is, in many ways, setting yourself up for failure. Users struggle to find content that matches their preferences, and the long tail of the catalog remains underutilized.
“Across different eras, the metadata was inconsistent in both style and emphasis.”
Rethinking metadata as a scalable, AI-driven layer
Plex quickly recognized that this was not a problem that could be solved incrementally.
Manual metadata processes could not deliver the consistency required. Nor could they scale to a catalog growing into tens of thousands of assets without becoming prohibitively expensive. More importantly, they could not capture the depth of understanding needed to power modern streaming experiences.
The team made a deliberate decision to move toward AI-generated metadata.
What they needed was not just automation, but a system capable of understanding content across multiple dimensions – not only what a title is, but how it feels, how it evolves, and how it connects to other content.
Vionlabs stood out by delivering exactly that.
Instead of relying on surface-level tagging, the platform analyzes content in depth, extracting structured metadata across genres, microgenres, keywords, and mood. It creates a consistent layer of understanding across all content, regardless of origin, era, or language.
This consistency became the foundation Plex needed to scale.
With a single processing pipeline, Plex could enrich metadata, power discovery, and unlock additional capabilities across the platform.
“We were looking for something that could deliver depth across multiple vectors, and we did not find anything else like it.”
From content processing to a multi-purpose capability
The value of this approach extended beyond metadata alone.
Plex could identify optimal ad break positions within each asset, allowing ads to be placed in moments that feel natural and non-disruptive, preserving the viewing experience rather than interrupting it.
On top of that, Vionlabs supports Plex with the generation of high-quality, engaging preview clips. This transforms the browsing experience from static to dynamic, helping users better understand and connect with content before they press play.
This meant that a single investment was not solving one isolated problem, but creating a shared capability that supports multiple parts of the platform.
Integration that fits the business, not the other way around
One of the defining aspects of the implementation was how naturally it fit into Plex’s existing infrastructure.
Rather than introducing external dependencies, Plex was able to run Vionlabs within its own environment. This allowed the team to maintain full control over processing while aligning with their existing workflows.
This flexibility proved essential.
Content ingestion is not constant. At times, large volumes of new assets need to be processed quickly. At other times, the system operates in a steady state. Being able to scale processing up or down based on demand allowed Plex to remain both responsive and efficient.
“We can scale the processes up or down depending on demand.”
Despite the sophistication of the technology, the integration itself was straightforward.
“It was actually really easy to integrate with our workflows.”
As the system became embedded, its impact expanded across the organization.
Product and engineering teams used it to improve recommendation systems. Editorial teams used it to uncover valuable content within the catalog. Marketing teams used it to identify opportunities for promotion. Ad sales teams used it to build more relevant and contextual offerings.
What emerged was not just a tool, but a shared layer of intelligence across the business.
“We’re now seeing the data used by a variety of teams to help grow the business.”
From better data to measurable business impact
The real value of improved content understanding is reflected in how users engage with the platform.
By incorporating richer metadata into its recommendation systems, Plex became significantly better at aligning content with user intent. Mood-based data played a particularly important role, enabling a more nuanced understanding of what users want to watch in a given moment.
The impact was measurable.
A/B testing showed a clear uplift in viewing minutes when mood data was used effectively. Users were not just finding content more easily, they were staying engaged for longer.
“We have seen significant uplift in minutes based on using mood data properly.”
At the same time, improvements in ad break placement helped protect the viewing experience.
Poorly timed ads can quickly reduce engagement. With a deeper understanding of content, Plex was able to place ads in moments that feel more natural, maintaining flow and minimizing disruption.
“Bad ad breaks can destroy engagement, and we have proven that Vionlabs improves it.”
Together, these improvements strengthened both the user experience and the business model.
ROI and long-term value
For Plex, the return on investment is clear and measurable.
Higher engagement leads to longer viewing sessions. Better ad experiences improve monetization. More efficient workflows reduce operational overhead.
Because Vionlabs is deeply integrated into the platform and applied across a continuously growing catalog, the value compounds over time.
“We know for a fact, based on the engagement we have seen, that it is ROI positive.”
Beyond immediate impact, there is also a strategic advantage.
By building a consistent and scalable layer of content understanding, Plex has created a foundation that supports future innovation across personalization, advertising, and content strategy.
“I do not know how we could have done it without Vionlabs.”
A partnership built on collaboration
From the beginning, the relationship between Plex and Vionlabs has been highly collaborative.
Rather than simply delivering a product, Vionlabs worked closely with the Plex team to identify the most valuable use cases and ensure that the technology delivered real business impact.
Over time, this collaboration has allowed the system to evolve alongside the platform.
New use cases have been introduced. Existing ones have been refined. And the role of content understanding within Plex has continued to expand.
“They helped drive us toward the use cases with the highest return.”
Building the foundation for the future of streaming
As streaming platforms continue to scale, the ability to understand content deeply and consistently becomes a defining advantage.
For Plex, that capability is now embedded in how the platform operates.
It shapes how content is discovered, how users engage, and how value is created across the business.
“Vionlabs enabled us to do the things we always wanted to do, but could not do at scale.”
And for users, it simply results in a better experience.
With Vionlabs, you won't.
With Vionlabs, you won't.
