Claudia Lalau Doesn't Choose Between Craft and AI
Claudia Lalau can spend four or five hours on a single shot. Then she opens social media and sees another creator release an entire film in a day.
“Come on,” she laughs.
But something about it disappoints her. The speed is real, as are the explosions, elaborate fight sequences, and 30-second clips produced from a single keyframe. But Claudia watches the details change between cuts. She notices when a character’s face shifts or if their grip on an object seems physically impossible. Beat by beat, frame by frame, she sees everything.
“The fight scene was amazing,” she says of one AI-generated sequence she saw online. “But if you look at every single shot, there is no consistency.” There lies the difference between a spectacle and a film. For Claudia, consistency is more than a technical standard. A character must remain recognizable long enough for an expression to build, a gesture to mean something, and most importantly, for an audience to feel it and care.
Her judgment is backed by two decades in production and post. Claudia started in animation, moved through editing, VFX, color, and finishing, then spent about a decade leading post-production in Dubai. There, she founded Bottomline Post-production, an operation that carried work from creative development through VFX, animation, editing, grading, sound, finishing, and final delivery, and remains active today. After moving to France, she opened Bottomline AI, expanding the studio into an end-to-end creative production and post-production partner that combines AI workflows with high-end traditional craft.
Her recent credits jump from Mercedes and Toyota to Bosch, Thermomix, Deliveroo, and BEEAH. As an active Sageworx Member, Claudia brought that experience to the agency’s JULABO video campaign. Sageworx sought her out for this particular client program because of the emotional quality of her filmmaking: she turns a generated character into someone worth caring about.
Claudia works where AI has to conform to a real production – a brief, a budget, a crew, a finished film. She directs fully AI generated and hybrid AI/live-action work, either leading the production end to end or folding it into an existing pipeline or live set. The tools have changed the scale of the work, but her standards have expanded with it.
I try again until I get the exact emotion I have in mind.
In previous lives, Claudia drew 2D animation frame by frame. She built characters through CGI, where movement required models, rigs, skeletons, textures, lighting, and material. AI has compressed much of that process, but it has not changed what she believes makes a character feel alive.
For What the Deep Knows, her beautiful animated short about a girl afraid of monsters, Claudia approached the central character in the same manner she would an actor. She asked for a specific reaction, studied the result, then tried again when the character gave her something emotionally wrong.
“It’s just as if I’m inspiring or directing a real person (with AI) to act like this, or move like this, or behave like this, to be sad, to react,” she says. “You treat it as if you are treating a human being.”
A dragon can make someone cry, she explains, same as a drawn child or a synthetic person, if the expression, gesture, and timing feel true. Claudia is the puppeteer, making choices until the character carries the feeling she imagined.
I’m very stubborn.
So is AI.
AI frequently returns something close to what you prompt for, but not quite. Some directors accept the surprise and build around it, but Claudia is driven by her vision, as opposed to the technology that makes it possible. She changes her mind only when the result earns it.
“I can see it already in front of me, this is how I want it, and I keep repeating until I get it,” she says. “But it’s exactly like a director on set. Sometimes they want a specific reaction or angle, and then when they shoot a different angle, they say, ‘No, I prefer this one.’”
She recently helped a learner fix an AI-generated image of a woman holding an oversized camera with an impossible grip. After 20 or 30 failed attempts, they stopped arguing with the still. They made the camera smaller in Photoshop, animated the image, and found the believable hand position in one of the new frames.
“She actually got the frame that she wanted from the animation, not from the still frame,” Claudia continues.
You don’t need to try every single platform.
Claudia tests new image, video, language, and audio models constantly, but if each novel release demanded her attention, she would get nothing done.
“Every single day there is a new platform, and it’s becoming annoying,” she says. Her advice is to choose one or two platforms that feel comfortable, then use them to evaluate new models for price, control, efficiency, and fit.
The shiny new tool may generate a longer, more realistic clip, but it could cost more – or make creative decisions somewhere in the middle. That’s why Claudia recommends an earlier version of Seedance: it gives her more control. Thirty seconds is a perk, but if those thirty seconds stray far from the original ask, the benefit disappears.
Her background as a university instructor also shapes how she helps other filmmakers make those choices. Claudia spent three years teaching Storytelling and Animation software, and now helps filmmakers choose tools that complement their own skills and specializations. On November 6th, she is scheduled to speak at the Athens AI Film Fest in Greece, where the theme this year is “Myths Meet Machine,” connecting ancient Greek mythology with the future of filmmaking.
I love the diversity in one project.
Claudia has spent her career coordinating teams across roles, cultures, and continents. In Dubai, she led post-production crews of 10 or 15 while liaising with agencies, directors, producers, and specialists around the world. AI has since compressed those crews to three, four, or five, giving her more creative responsibility, which she prefers, but that is not the same thing as working alone.
At Sageworx, her direction sits within an agency relationship that begins before she joins the production and continues after her shots are finished. For JULABO, that work began with a storytelling problem. Its temperature-control equipment operates in highly technical settings, often out of sight. Sageworx wanted to display its relevance through scenes and locations the client could not film conventionally. AI opened up those possibilities, but the agency still needed to decide what was worth showing.
The initial approach was cautious, since machinery, close-ups, and hands, with no faces or human performances that might expose the limits of AI. Claudia kept pushing for a human “heartbeat.” As people entered the story, the team had to judge their performances as closely as the technical realism. After all, a character could look convincing and still carry the wrong feeling for the project.
Her reaction was driven by the same stubbornness that keeps her working on a single shot for hours at a time: Claudia can see the feeling she wants, and she has the patience to keep working until it appears. Sometimes, she admits, she pushes the emotion beyond what the brand can hold. In one early version, the central character’s sadness overwhelmed the story. The Sageworx team helped redirect the performance toward someone eager to participate in life but frustrated by a physical inability. That gave Claudia a more specific emotion to pursue, which is her preferred form of feedback.
The exchange captures what Claudia values in a creative partnership: freedom to push, honest feedback, and the right chemistry within the team. She feels “the energy around Sageworx is amazing.”
Going forward: more narrative.
This November, in Prague, Claudia joins Calibration, an arthouse psychological drama from director Serani Hunyi, adapted from L.B. Schulman’s award-winning short screenplay. The film blends live-action performance with AI environments set in Prague, 2040.
“After twenty years in advertising, I felt like storytelling is what I actually came to this industry to do,” she says. “Calibration is the start of it, not a one-off.”
The feature is one part of that move. Claudia has also been named an ambassador for the World AI Film Festival WAIFF and is developing an original animated series for children. Her personal series, Nick and Meeva, is inspired by her twins, Nicholas and Maëva, who now ask when the next episode is coming.
They think they are stars,” she says. "And they are.
Claudia’s looking forward to the next generation – not just for the input of her twins, but the teenagers who are old enough to make films of their own.
On September 3 and 4, she held the first TeenAIgens Academy workshop in Beirut. From there, she wants to bring the program to France and across Europe, working with schools, festivals, cultural venues, and supporters who can help keep it accessible.
The curriculum comes directly from production. Every problem Claudia solves on a commercial, animation, or feature becomes something she can teach. Her hope is that they leave the workshop knowing how to think with AI, use it responsibly, and understand how much work sits behind the promise that one prompt can make a film.
She has already heard the objection from teenagers: “Why do I need this workshop? I can just tell Gemini to make a film for me.”
That misconception inspired her curriculum.
“We need to be preparing them now,” Claudia says. They have access to so many generative tools and platforms. Teaching them to leverage technology without giving up the director's chair is important to their future as a filmmaker.
They’re not going to be directors on the second day.
The models will keep changing. The newest one will make longer clips, sharper faces, and bigger promises. Claudia will still be watching the person inside for the detail that makes us care about them, and teaching the next generation of filmmakers to watch for it, too.
Follow Claudia’s work on LinkedIn, Instagram, or YouTube, and connect with her through Sageworx.
Claudia is one of 100+ senior specialists in the Sageworx field. Once a week we send out what's moving through it.