You Don't Have to Cut the Big Scene
The Gigaclear film needed a village in rural England, a hangar with a stage, and a seated crowd of 100 with a path down the middle. None of it existed. It was delivered as a national UK TV commercial in about four weeks, through M3 Labs.
Send the brief →A rough brief is enough. You get a reply within a day.
UK full-fibre broadband
Mar to Apr 2026
The finished Gigaclear TV commercial, national UK campaign
What the client needed
The same job the ad has always done
Gigaclear's commercial had the job their advertising has done for years: reach the people it is meant for, deliver one clear message, and protect the brand. None of that changed because AI was involved.
What changed was how far the idea could go. On the same budget as before, the film could be bolder: a village in rural England, a hangar with a stage, and a seated crowd of 100 with a path down the middle. That is a world this budget could not reach before.
What had to go right
Every shot has to agree with the last one
AI can make a single picture of a hangar easily. The trouble starts with the second picture. Ask for the same hangar from another angle and the room comes back a different shape. The crowd comes back with a different number of people, in different seats, with different faces. A commercial is many angles of one place, and pictures that disagree with each other cannot be cut together.
The result
A bolder film that still does its job
The film was delivered through M3 Labs as a national UK TV commercial, in about four weeks. It carries the message it was written to carry, and it has these in it:
- A house, a street and an attic room that read as one place in rural England
- A hangar that is the same room from every camera angle
- A seated crowd of 100 that keeps its layout in every shot
- Filmed actors inside sets that were never built
Why it holds together
Each place was built once
The hangar and its crowd were built once, as one space. Every shot of them is a different camera position in that same space, so the shots agree with each other.
Send me the script or the storyboard. Within a day you get a written reply: what is realistic and which way of working fits. This first read is free.
Send the brief →how it was done
That was the result. What follows is the full technical breakdown: how the locations, the hangar, the crowd and the window shot were made, the tools, and the options that were offered. If you only needed to know whether a scene like this can be done, you have already read that part.
▶ Watch the case study · 3:11
How the commercial was made, in three minutes: locations, the 3D hangar, the 100-person crowd, real actors and the window shot
What was not generated
- The actors' performances were filmed.
- The hangar layout, camera positions and lenses were set in Cinema 4D.
- The astronaut, the hero of the film, changed with every generation, so he was rebuilt as a rigged 3D character.
- The crowd was placed in the scene by a script.
- The edit, the compositing and the finishing were done by hand.
// THE TRAP OF THE FINISHED LOOK
AI can make something look finished before it is even usable.
That was the trap.
The real work on the Gigaclear commercial was turning beautiful, unstable AI output into controlled production footage.
The Gigaclear national UK TV commercial (2026, produced through M3 Labs) was made with AI-generated locations and characters. A Cinema 4D build of the hangar, a scripted 100-person crowd and real actors composited into the generated sets kept it steady from shot to shot. Kevin Kuteli (Wobbleman) was the creative technologist. The performances, the edit and the finishing were done by people.
This was a full commercial production, with a director, producer, agency and client, and with notes arriving while the project was already moving. AI was central to the process, and it changed where the production work happened.
I came into the project during the second week. There was a storyboard and an early look, so the tone and a rough idea of the locations were already there. My job was to take that early direction and turn it into shots a commercial could use.
01 · Locations
The AI locations: rural England, generated and art-directed
On my first day I started producing more location options from the visual direction the director had already provided. The setting had to pass for rural England, so the generated locations still had to be believable.
That meant research. We looked at how houses in rural England look inside and outside: window shapes, the number of windows, sidewalks, streets, house spacing, exterior materials and interior details. A generic AI village would have been easy to get. We needed a place that felt real enough for people to understand the world of the commercial.
The main locations were the interior and exterior of the house, the street outside, the hangar interior, and the attic room. Each one had to work as a production space. The house interior had to feel lived in, the street had to belong to that specific house, and the hangar had to hold the stage, the presenter, the speech and the crowd.
During that same week, we were also working with the agency and client to lock down exactly how those locations should look. Each location was taken through agency and client notes until it was approved: generated, then retouched and composited by hand where a note asked for something specific.
When you generate images for yourself, you can accept whatever looks best. For a director, an agency and a client, you are building toward a specific creative and production requirement.
In traditional production, once you choose a location, many decisions are already made for you. The windows and walls are where they are, and the room has a size. You can dress it, light it and shoot it differently, but the physical location gives you boundaries.
With AI, those boundaries become flexible. You can move a window, change a street, rebuild a room, adjust the time of day, or create a location that would be difficult to find in real life. That freedom also creates more decisions. If everything can change, everything can become a note. I have written separately about why "change one thing" is the hard request.
So this stage was location scouting, art direction and production design at the same time, done with AI.
02 · Storyboard
From storyboard to production frames
Once the locations were approved, the next step was creating frames from the storyboard. This meant producing different angles, working out character positions, adjusting object placement, testing lighting moods, and figuring out how the storyboard could live inside the AI-generated locations.
AI breaks down here if one prompt is asked to do everything. A model cannot hold the storyboard, the approved location, the character, the shot size, the lighting, the edit and the director's intention all at once. It can give you a frame that is mostly right, where the part that is wrong is the part that matters.
For that reason, I worked across different platforms and workflows. The agency's main platform was Freepik / Magnific, which became one of the core places for generations and structured spaces. I also used Martini, which was useful at the time for 3D Gaussian splatting and creating scenes from images. I used ComfyUI locally to test angle workflows, reference-based generations, and different ways of exploring how the shots could match the storyboard. On top of that, I used Photoshop, After Effects, Cinema 4D, and custom scripts whenever the AI tools alone could not solve the problem.
The tool changes with the problem. One shot needs generation, another needs retouching, compositing, 3D or a script, and some have to be rebuilt a different way because the model will not give the control the shot needs.
One of the workflows I built turns storyboard frames into more developed production frames. It takes the storyboard, the animation direction, the visual style and the approved location language, and uses them to create frames much closer to the shots we needed.
The workflow did not produce final frames. It gave us a way to test angles, see how the storyboard could sit inside the AI locations, and show the director and agency what the shots could become. Each frame was then finished by hand: retouching, objects moved, lighting adjusted, and parts rebuilt where the AI had misread the space or changed something already approved. There is a longer piece on how characters and products are kept consistent.
AI is very good at a fast, polished-looking base, and a concept frame can look close to finished to someone outside the process. A commercial-ready shot has to match the story, the lighting, the camera, the edit, the client's notes and the surrounding shots. That last stage is where the production control happens.
03 · Live action
Compositing real actors into AI sets
The actors were filmed, and the locations were generated. The AI image had to sit next to real footage and feel like part of the same world.
There were calls with the director and DOP about lighting, time of day, and how the footage would connect to the AI environments. If an actor is shot in real life and the environment is generated, you have to decide where the light is coming from, how strong the shadows are, how the color temperature feels, and how much room the shot leaves for changes in post.
There were two approaches. Where we were confident in the final look, we could match the lighting from the beginning and bake the mood, shadows and direction into the AI environment. That usually gives a more realistic result, and it makes changes harder if the edit or client direction shifts later.
The other approach was to keep the lighting more neutral, with softer shadows and less specific direction, so there was more room to adjust later. The frame can then feel less dramatic at the concept stage.
We chose between the two shot by shot. There were also specific visual problems, like the flying letters and paper around the character, where the real footage and AI-generated elements had to work together. At that point the job is VFX, compositing and post-production again. I have set out the VFX work an AI shot still needs in its own article.
04 · The hangar
The hangar: why it was rebuilt in 3D
By the third week, more artists had joined and we were working with the director and client to define final frames, lighting, time of day and the feeling of the shots. I also started testing animation sequences to show the director, so we could judge timing and see whether shots needed to be shorter, longer, added or removed. This is when we built the hangar in 3D.
The hangar was the main space of the film: where the Giga hero element would hold the speech, where the crowd would sit, where the stage would be placed, and where several important camera angles had to work. With prompting alone, the space would have kept changing. The camera angles and the layout would not have held, and the director would not have had a stable space to make decisions in.
In the 3D hangar we could plan camera angles, test lenses, think about depth of field and movement, place the stage and the character, and discuss shots with the director while screen-sharing the scene. We could move the camera around live and make practical changes before going deeper into final animation.
The build gave structure to a process where AI could otherwise become unstable. Prompt-only workflows tend to skip this step, because it takes technical skill and production planning.
When AI could not give us reliable spatial continuity, the solution was to go back to production fundamentals: camera, lens, scale, blocking, staging, lighting, and layout.
05 · The crowd
The 100-person crowd: generated in rows, placed by script
We needed exactly 100 crowd characters inside the hangar, sitting in a specific arrangement. The crowd had to be divided into two sides with a middle pathway for the presenter to walk through. The positions had to feel locked, the layout had to work across multiple camera angles, and there could not be obvious repeated people or repeated faces. It had to look staged, the way a production would stage it.
AI is bad at this. It can generate the idea of a crowd, but it cannot hold a full crowd of specific figures in exact positions, with consistent spacing, across multiple angles.
We also did not have time to solve it the long traditional way, where every person is built as a full 3D character or handled through a much heavier crowd pipeline. It needed a faster route that still gave a strong production result.
The workflow I built makes crowd rows instead of individual people. I generated groups of five to seven characters at a time using Freepik / Magnific, then created front and back views that could be used to build AI-assisted 3D-like crowd elements. One character per chair was the other option, with more time. Rows reached the quality the shot needed sooner.
I then wrote a script to bring those crowd elements into Cinema 4D and place them. It replaced importing, aligning and spacing every element by hand. When I opened the file, the two sides of the audience were already there, with the central pathway between them.
That gave us a controlled base: a crowd designed around the path, the camera and the staging.
The problem was simple to describe, and no existing tool solved it. The solution was built from pieces: AI generation, front and back views, AI-assisted 3D, Cinema 4D layout and scripting. Many of the important problems in AI production are like this. You have to understand the creative goal and the technical limit, and build a practical path between them.
06 · Two routes
A live-action crowd shoot compared with the AI and 3D hybrid
For the crowd scene, a practical shoot and an AI / 3D hybrid production need very different things. This is how the two routes compare.
| Category | Traditional Live-Action Shoot | AI / 3D Hybrid Production |
|---|---|---|
| Crowd | 100 real crowd actors / extras | 100 digital crowd characters |
| Location | Real hangar rental, permits, insurance | 3D hangar build and digital scene setup |
| Crew | Director, producer, DOP, camera, lighting, art department, wardrobe, makeup, production team | Creative technologist, AI artists, 3D artist, compositor, animator |
| Main Cost | Crew, location, equipment, crowd actors, logistics | 3D setup, crowd workflow, AI generation, animation, compositing |
| Main Challenge | Coordinating a large physical shoot | Keeping space, crowd, camera, and animation consistent |
| Flexibility | Lower after the shoot is done | Higher, but revisions can still be technically heavy |
| Best Summary | Physically real, with a large shoot to organise | Smaller footprint, still production-heavy |
The hybrid route was still a lot of work. What it allowed was a much bigger idea with a smaller footprint. AI does not make production free. It can put certain ambitious ideas within reach, if the team knows how to control the process.
07 · Animation
AI animation and the window shot
In the fourth week we moved into animation. I used a mix of Cinema 4D renders, storyboard frames, AI video generation, prompt editing, video-to-video workflows, extensions and post-production cleanup. Each shot got the method it needed: some straight from a Cinema 4D render and a storyboard frame, others through video-to-video and cleanup.
One shot can work quickly and the next can break for reasons that are not obvious at first. The model may hold the movement and lose the environment, or keep the environment and change the character. A shot can also look right until it is in the edit, where the timing does not work.
The window shot
Near the beginning of the commercial, the director wanted a specific camera movement: through the exterior and in through one window.
There were two routes. One was a simple box model of the house in Cinema 4D: move the camera through primitive geometry, then use the first and last frames as references for video generation. The other was a detailed prompt with references.
The prompt and references held the move, so the question became which window. The director was shown the side windows and the slit window in the middle of the shot, and the slit window was chosen.
The shot was not done at that point. As the camera moved toward the real performance of the actor, the environment started to hallucinate. That was fixed with video-to-video work, cleanup, regeneration and video extension, so the shot would fit the final edit.
You do not always know what is missing until the footage is in the edit. A shot can look good by itself, and once it sits between the previous and next shot you may see that it needs a few more frames, a cleaner transition, a better ending or a different movement. That is where post-production experience matters: you make what the edit needs.
08 · Result
What the job shows about AI in commercial production
The problems along the way were ones I expected. Consistency, specific client notes and unpredictable animation are the failure points I have worked around in post-production for more than a decade, and I planned for them.
AI made pre-production faster and let the team attempt a world that would have been hard to find or build as a physical set. The finishing still took production time. A first version of a shot appears so quickly that the rest looks as if it should be quick too, and that is the hardest part to explain on a job like this. A concept frame and a finished shot are different pieces of work. Here the distance between them was covered by a 3D hangar, a scripted crowd, compositing and an edit, done by people who knew what to do when a model broke. The same order of work is set out in the seven stages of an AI commercial.
Questions
Gigaclear AI commercial: questions and answers
Was the Gigaclear advert made with AI?
Yes, with people in charge of it. The locations and characters were generated with AI. The finished commercial also needed human direction, a 3D build of the hangar, live-action footage of the actors, compositing, editing and finishing.
Which parts of the Gigaclear commercial were not AI?
The actors' performances were filmed. The hangar layout, camera positions and lenses were set in Cinema 4D. The astronaut was rebuilt as a rigged 3D character. The crowd was placed in the scene by script, and the edit, compositing and finishing were done by hand.
How was the 100-person crowd made?
The crowd was generated in rows of five to seven people, with front and back views of each row. Those rows were turned into 3D elements, and a script placed them in the Cinema 4D hangar on two sides of a central pathway, so the layout held across camera angles.
How were real actors combined with the AI sets?
The actors were filmed separately. Before the shoot, a 3D stand-in was placed inside the AI set to check position and scale, and lighting and time of day were agreed with the director and DOP. After the shoot the footage was tracked, composited into the generated environments and cleaned up.
How long did the Gigaclear commercial take?
About a month, from the first week of story and storyboard work to animation and finishing in the fourth week.
Whether you are an agency, a brand or hiring me directly, send the brief, script or storyboard and say how you will measure success. You get a reply within a day: what is realistic and how the production should run.
Send the brief → See how I work →Kevin Kuteli
AI Creative Director & Creative Technologist. Came to AI through production, not prompting: an editor, motion graphics lead and VFX supervisor before the first model. Works with agencies, directors and brands as Wobbleman.
// Written from direct production experience on the Gigaclear national commercial.
← Back to Work