J'ai Essayé de Créer une Vidéo à 400'000€ avec l'IA

The presenter begins by recounting a significant production from three months prior: a 4-minute, 30-second advertisement that cost €41,000. This elaborate production involved 22 hours of filming, 15 extras, over 10 technical crew members, and unique elements like a rented crocodile, a Lamborghini SVJ, and a fake wedding. Despite its success in garnering over a million views, the presenter now believes he would not invest such a sum again, given the rapid advancements in Artificial Intelligence (AI). The core of the video is an experiment to recreate this €41,000 advertisement using 100% AI, with zero actors, cameras, or sets, aiming to spend only a few hundred euros. The presenter estimates that an equivalent real-life production of the AI-generated video would cost over €200,000, but he managed to create it for less than €100 in a single afternoon from his office. The promise to the audience is to reveal the exact process and provide a plan to monetize this emerging AI video creation skill before others. The production process is broken down into five steps: 1. **Script and Storyboard:** Reworking the original ad's text, brainstorming scenes, and seeking inspiration for visually impactful sequences. 2. **Generation:** Using the XField platform to generate each sequence with AI prompts, iterating until perfection and coherence are achieved. 3. **Assembly:** Combining all elements: editing, voiceover, color grading, and music. 4. **Comparison:** A side-by-side comparison of the AI version with the original €41,000 advertisement. 5. **Monetization:** Providing a detailed plan to start earning money with this new skill. The original advertisement's concept revolves around Stéphane, an office worker on the verge of burnout. He quits his job, discovers e-commerce through the presenter's video, joins the "Ecom Accélérator" program (promising €10,000 in three months or a full refund), works intensely, achieves his goals, and ultimately enjoys a luxurious life with a Bugatti. For the AI recreation, the presenter shortens the original script from 4 minutes to 1.5-2 minutes. He uses an AI tool, Cloud CowerK, to condense the script and then gathers visual inspiration from Instagram and YouTube for AI-generated content. The revised storyboard includes dramatic scenes like Stéphane's office exploding, him burning his workplace with a flamethrower, and eventually transforming his shabby apartment into a luxurious penthouse, complete with a Bugatti crashing through a window. The generation phase begins with XField, specifically its Cinema Studio 2.5 product, chosen for its advanced video generation capabilities and integration of various AI models. A key feature of Cinema Studio is the ability to create and save character and location references, ensuring visual consistency across different clips. The presenter demonstrates this by uploading 50 photos of himself to create an AI character, then generating a reference sheet from various angles using Nano Banana, an integrated XField tool. The same process is applied to create Stéphane, the main character, and various locations like the open-plan office, Parisian metro, and penthouse. The presenter then proceeds to create individual clips, detailing the prompts and AI tools used. For instance, the first clip, Stéphane's arrival at the office, uses a "single shot" mode in Cinema Studio, with a starting image of a desk and an ending image of Stéphane. The prompt describes a super-fast tracking shot through the open space, stopping abruptly on Stéphane. The second clip, showing "time passing," uses Nano Banana Pro to age Stéphane in the final image, then animates the transition with a time-lapse effect in Cinema Studio. Further clips showcase increasingly complex and absurd scenarios, such as Stéphane literally sinking into his office floor, being "saved" by the presenter, a close-up of Stéphane's rage, his destructive rampage with a flamethrower, and a triumphant exit from his burning office. A unique clip involves a drone shot over Paris, showing thousands of pay stubs falling like rain, effectively rendered using Sora 2 (another XField integrated tool) for its superior text rendering capabilities. The narrative continues with Stéphane discovering the ECOM Accélérator program on his phone in the metro, making the purchase (leaving him with only €5,000 to emphasize his commitment), and embarking on an intense work period. Clips depict his apartment becoming a "battlefield" of pizza boxes and soda cans, animals randomly appearing in his workspace (a parrot, a raccoon, an ostrich), and him working with multiple arms, like a Hindu deity, or walking two crocodiles through Paris (a nod to the original ad's crocodile). The turning point comes with Stéphane's first sale, visualized by a subtle shift from cold to warm lighting and a green Shopify logo reflecting in his eyes. This leads to euphoric celebration, and then a significant "poverty to wealth" transition, morphing his humble apartment into a luxurious penthouse, followed by him purchasing and driving a Bugatti, even dramatically crashing it through his penthouse window. The final scene recreates the original ad's ending, with the presenter surrounded by multiple crocodiles, delivering a call to action. After generating all clips, the presenter adds a voiceover (which he records himself, though AI could have done it) to narrate Stéphane's story. The finished AI-generated advertisement is then presented, showcasing its narrative flow and visual impact. A direct comparison is made: the original ad cost €41,000, involved 22 hours of shooting, a 25-person crew, 15 extras, and various expensive props. The AI version cost €100 for XField subscriptions and generation credits, took 4 hours of work, and was done by one person. An AI (Claude) estimated the real-world cost of producing the AI version without AI tools to be between €200,000 and €400,000, highlighting the massive cost and time savings. The presenter emphasizes that AI not only reduces costs but also enables the creation of scenes previously impossible or prohibitively expensive to film. The video then shifts to the immense monetization opportunities. The presenter explains the "chasm" phenomenon, where early adopters of new technology gain a massive advantage. He cites statistics: 86% of US advertisers already use or plan to use AI for video ads, and by mid-2026, 40% of all digital ads will be AI-generated. The video ad market is over $200 billion annually, with many businesses citing high cost and slowness as barriers to video marketing. AI offers a solution, producing premium, cinematic content for a fraction of the traditional price and time. Four main monetization methods are outlined: 1. **AI Creative Agency for Brands and Businesses:** Producing viral ads for brands at a significantly lower cost and faster turnaround than traditional agencies. The presenter suggests finding clients through freelance platforms or direct outreach, offering free improved versions of their existing ads as a demonstration. 2. **Selling a Course/Training:** Becoming an expert in AI video generation and offering a paid community or course. The presenter mentions a US example earning $140,000 per month with a community of 1,400 members at $97 each. This market is largely untapped in France. 3. **For One's Own Business:** Leveraging AI for e-commerce, services, or physical businesses to create product visuals, generate multiple ad variants, produce organic content, and even create AI avatars to replace expensive UGC creators. 4. **Freestyle/Novel Applications:** Exploring unique and creative uses for AI video, such as cinematic virtual tours for real estate agents, MasterChef-style food videos for restaurants, dramatic transformation videos for fitness coaches, or fantastical wedding videos. The presenter concludes by warning about the impact of AI on traditional creative jobs (photographers, graphic designers, video editors), citing significant drops in job postings. He stresses that those who refuse to adopt AI will be left behind, comparing it to past technological revolutions like the internet or e-commerce. He acknowledges that AI is not perfect and requires iteration and practice but emphasizes that the first results are already impressive. He encourages viewers to try XField, promising a document with all prompts and settings used, and invites them to share their creations. He reiterates that passing to action now is the key to benefiting from this technological revolution.