AI Only When It Matters

In the media industry, AI has rapidly transitioned from a buzzword to an operational reality. While many companies are racing to adopt the technology, at Simply.TV, we focus on defining a principle for its use: AI must be reliable and compliment human expertise, not replace it.

“AI for the sake of AI does not make sense,” says Morten Trolle, Simply.TV’s CTO. He goes on to add,  “We are using AI because we genuinely see that it creates true value, backed by data.”   This perspective forms the foundation of Simply.TV’s strategy, which prioritizes practical value and measurable quality over trend-following. In other words, Simply.TV will only use AI when it creates value for the business that can be proven as a whole. 

To lead this venture and integration, Simply.TV has appointed Anton Golles as the new Head of AI. 

Testing, Testing and More Testing

Anton brings a mindset shaped by computational physics and a practical track record of building AI from generative animation in 3d to political research agents. His focus has been on designing self-improving agentic systems and moving those from prototype to production, a process centered on metrics and rigorous evaluation. 

This practical experience aligns with the disciplined approach at Simply.TV. Here, Anton is implementing a strategy governed by a straightforward playbook: first, question if AI is even needed. If the answer is yes, then question every requirement. Finally, the system is subjected to relentless testing until it is both unbreakable and matches Simply’s high editorial standards. 

For Anton, the appeal of the role lies in three key areas: the autonomy to design and execute the AI strategy, a culture he describes as a “tech-first meritocracy where the best ideas win,” and the constant challenge of learning within a new domain. “There’s a real focus on fun here, too. That balance is pretty unique,” he notes. 

His philosophy remains clear and grounded. “Our go-to approach is not to use AI. Only when we absolutely have to,” he states. “When we are building AI solutions, we always want to make sure it’s reliable in delivering the high-quality our customers expect.” The outcome is a careful integration where AI accelerates growth, while expert human verification ensures all automated outputs remain accurate and trustworthy.

Integrating AI, Not Overhauling Systems

The strategy for AI at Simply.TV is one of enhancement, not replacement. The company is thoughtfully weaving AI into its proven editorial workflows rather than rebuilding from the ground up. 

In practice, AI handles the initial, high-volume repetitive tasks such as drafting program descriptions or identifying languages. This allows human editors to focus their time and attention on review and refinement. These editors apply cultural nuance, ensure factual accuracy, and control the final data delivered to clients, guaranteeing the company’s quality standards are met. 

This collaborative model enables Simply.TV to scale its metadata operations efficiently as it expands globally, while maintaining rigorous editorial oversight. As Morten summarizes, “Ultimately, quality comes from our expert editors. It’s still the editors we listen to.” The result is metadata that is not only faster to produce but remains accurate, nuanced, and trustworthy. 

Current Applications: Deduplication

Simply.TV is currently applying AI in targeted areas where it can provide clear, verifiable support, such as managing complex content libraries. A prime example is the system for identifying duplicate program entries, which is a significant challenge as Simply continues to expand its global footprint. Duplicates never appear identical. Titles may be translated or transliterated, cast lists can differ, and descriptions come from various sources. 

To solve this, Simply.TV is implementing a multi-layered approach. The system will retrieve similar entries in the database, by assessing semantic similarity of descriptions and cast-list overlap, amongst other important features. This allows the system to find duplicates that heuristic-based systems do not discover.

These combined signals generate a confidence score for each potential match. High-confidence decisions are handled automatically, while cases falling into an ambiguous middle ground are sent for a deeper, contextual review. Here, a large language model (LLM) acts as an impartial judge, examining the full metadata of both entries side-by-side to make a nuanced final determination such as distinguishing a remake from an original. 

All candidates flagged by this intelligent pipeline are ultimately presented to editorial staff for final validation. This process ensures catalog consistency and accuracy, leveraging AI’s computational power for scale and pattern recognition while keeping expert human judgement firmly in the loop for final quality control. 

The Future

Looking ahead, we are exploring how AI can extend high-quality metadata standards from popular content to lesser-known titles. By training systems on our best editor-curated data, we aim to enrich metadata for a broader range of programs within a catalog. This levels the playing field for all content no matter where it’s from - big or small. 

Another development is Semantic Search, an upcoming feature designed to interpret natural language queries. This would allow users to search for content using everyday conversational phrases, with AI mapping the intent to Simply.TV’s metadata to improve discovery across platforms. 

In the end, the overarching theme of Simply.TV’s AI development is the application of the technology as a strategically calibrated tool. Our mission, now led by Anton Golles, is characterized by an emphasis on testing, measurable outcomes, and the preservation of editorial oversight. As we continue to scale with AI, quality remains at the forefront. Customers can expect faster results with the same quality Simply is known for. 

About Anton Golles, Head of AI 

Before joining Simply, Anton designed and built advanced agentic AI systems for business applications and developed data solutions to simplify existing workflows. He will apply this test-driven discipline at Simply.TV to ensure AI integration enhances scale and accuracy while firmly supporting editorial quality. Outside of work, Anton enjoys carving wooden artifacts and playing DnD with his friends. 

About Morten Trolle, CTO and Co-Founder

Guiding the company’s technological vision, Morten focuses on building state-of-the-art solutions to serve as the preferred metadata partner for content owners, broadcasters, TV operators and online media platforms. His leadership ensures the platform’s end-to-end automation delivers near-to-live TV listings, while a modern editorial approach enhances the data with high quality. When not building Simply.TV’s tech stack, you can find Morten watching his favorite show of all time - The Simpsons.