Media must stop playing with AI and start creating value

During the period when we launched the D-Smart project at the Doğan Group, I realized that becoming a digital platform would be possible not only by producing content but by making sense of the data at hand. While organizing archives with systems like Synegy, the ability to open raw footage from news reporters for editing on an ordinary computer led me toward a different responsibility. We built a system where we could watch live how people behaved, where they clicked, and at what moment they exited, even within the free live internet content provided by Star TV and Kanal D, which were under my responsibility.

Thanks to this system I installed in the office of İrfan Şahin, who was the general manager of Kanal D at the time, we were able to have an idea not only about the content but also about the impact of the content. Together with Mehmet Ali Birand, we positioned internet-related content at the end of the evening news bulletins like an "attach." By looking at the reactions, we rethought not just the content, but even the format and the order.

The year is 2025. The journey we started back then has now evolved into the phase of AI-supported hyper-personalization and data-driven content production. However, most of the media sector still uses AI like a decorative ornament. Having AI make a trailer or generate a few headlines won't move anyone forward. The issue is embedding AI into the system.

INTELLIGENCE IS NOT ENOUGH, DATA IS REQUIRED

It is no longer important how many people watched a series, but at which second who left the platform. More importantly, why they left. Was the scene boring? Did the character seem repulsive? Did the rhythm break? To answer these questions, you need an effective data architecture, not just intelligence.

It is essential that data is not just at the level of "how many people watched," but at a level that performs behavioral analysis on a scene-by-scene basis. The audience is no longer "a mass," but intersecting individual clusters. And to reach these individuals, metadata extracted on a second-by-second basis is required. In other words, which emotion is rising within the scene, which character is coming to the fore, which location is used, what kind of music is heard — all these elements must be tagged second by second.

NOT A MEDIA COMPANY, BUT A DATA COMPANY

This data transforms not only content production but also marketing campaigns. Today, it is possible to produce 20-second micro-trailers from a multi-episode series, tailored to different cultures and segments. Moreover, it is not humans doing this, but AI systems working based on data. Thus, even marketing content is fed by the content itself.

Traditional media giants can trust their content archives and creative staff. But if that content is not processed, it only gathers digital dust. That is why media companies must now redefine themselves not as content producers, but as creative platforms that process data.

TRYING IS NO LONGER ENOUGH

Companies that experiment with AI are only remembered as the "first to do it." But those who integrate it into their processes create a lasting impact. The era of "we use AI too" is over. Now, the era of "we earn more thanks to AI" has begun.

And the question that needs to be asked in this era is this:

It is not how smart your AI is — it is how ready your data is!