A title sequence made entirely with AI

• ClientAlmashhad TV
• Year2024
• RoleAI director · Motion designer
• ToolsGenerative AI pipeline — end to end
• DisciplineTitle sequence · Arabic calligraphy
• Traditional 3D usedNone

The experiment

The full title sequence for Menna Wa Fena Season 2 — luxury gold Arabic calligraphy on black marble — was produced entirely through generative AI pipelines. Every frame, every texture, every light reflection. No Blender, no Cinema-style 3D render, no traditional compositing chain.

That was a deliberate test, not a shortcut. The question was whether a generative pipeline could hold a consistent premium look across a complete broadcast title sequence, or whether it would fall apart the way AI video usually does the moment you need shot-to-shot continuity.

What actually worked

• Material and light. Marble, gold, caustics, soft specular rolloff — generative models are extremely good at this, and getting the same look out of a path tracer would have cost days of render.
• Atmosphere. Dust, haze, depth falloff came free rather than as a comp layer.
• Iteration speed. Directional changes that would be a re-render became a re-prompt.

What didn't, and how it was handled

• Arabic letterforms. Generative models still mangle Arabic script — connections break, diacritics invent themselves. Type had to be driven and controlled deliberately rather than trusted to the model.
• Continuity. Holding one marble surface and one gold treatment across a whole sequence takes reference discipline and a lot of rejected output.
• Timing. Models produce clips, not edits. The rhythm of the sequence is still a human decision, made in the cut.

The honest conclusion

Full-AI production is viable today for texture-led, atmosphere-led, non-typographic sequences — and only if someone in the chain has traditional motion design instincts about pacing and restraint. It is not yet viable when precise brand geometry or Arabic type has to be exactly right. Which is why the cinematic Arabic title open from the same year is a hybrid: Blender 3D for the type and structure, AI for smoke, embers and atmosphere.

Both approaches are in my pipeline. Choosing between them is the actual skill. More on how I use AI video →

Next: Andi Soual — award-winning visual identity →