Que vaut l’IA de YouTube Studio ? (J’ai peur)

This video explores the capabilities of a new AI chatbot integrated into YouTube Studio, designed to help creators improve their channel performance. The presenter, initially concerned about low views on his latest video, discovers this feature and decides to test its potential. The AI chatbot, referred to as "Studio" or "IA," has access to comments, video content, and channel performance data. The presenter begins by asking it to help improve his recent video's performance. The AI notes that the video's performance is atypical compared to the channel's usual trends. The presenter is impressed, suggesting that YouTube has integrated more than just a language model like Gemini; it has built in expertise on YouTube strategy. He humorously questions if this makes his role as a coach obsolete. The presenter then delves into testing the AI's analytical abilities. He asks it to identify his channel's best-performing topics and formats based on the last 100 videos. The AI identifies "YouTube and social media ecosystem" and "personal development and psychology" (specifically ADHD and autism-related content) as top themes. Crucially, the AI goes beyond simple view counts, analyzing audience comments to highlight that honesty in his videos boosts watch time. Next, he focuses on click-through rates (CTR) for his last 50 videos. The AI compares titles with the best and worst performance, listing successful titles like those related to making money, autism/ADHD, and tool testing. It also identifies poorly performing titles as being too academic or vague. The AI offers advice for future titles, suggesting angles like "the truth about X" over "how-to" formats, though the presenter notes that "how-to" titles can perform well with the right audience. The AI also emphasizes using personal experience, suggesting titles like "how I hacked my freedom" instead of "how to become free." The presenter finds these title suggestions to be good, recognizing they are tailored to YouTube. A significant portion of the video addresses the effectiveness of Shorts versus long-form videos for audience building. Based on the last 90 days, the AI unequivocally states that long-form videos are superior, generating over 3000 new subscribers compared to 187 from Shorts, a 16:1 ratio. Shorts act as an "exposure window," with 74% of viewers not knowing the channel. The presenter then expands the timeframe to 365 days, finding that long-form videos still dominate, with Shorts generating nearly a million views but only 1400 subscribers. The cost per subscriber is significantly higher for Shorts (670 views vs. 100 views for long-form), making long-form videos six times more effective for subscriber acquisition. The AI concludes that Shorts are good for awareness ("top of funnel") but business and loyalty are built on long-form content. The presenter then explores the AI's ability to analyze comments, a task he finds time-consuming and distracting. He asks for the three most frequent topics or problems mentioned in the comments of his last 10 videos. The AI identifies: 1) neurodivergence as a key for understanding, specifically the difficulty adults face in getting diagnosed; 2) the "AI rat race" and fear of obsolescence or losing authenticity; and 3) the struggles of small creators, referencing a specific video. It also notes a recurring humorous mention of his "mustache" as a sign of audience loyalty. The AI suggests a video on organizing work with ADHD, bridging his psychology and business themes. He then probes for specific questions his audience hasn't had answered. The AI highlights questions about his Trello organization, tools for managing mental load, and external brain methods, suggesting a demo video for his tools. It also identifies audience interest in succeeding on YouTube in 2026 without fast cuts or how to find quality closers. A key insight is the AI's analysis of audience preference between business and lifestyle/psychology content. It concludes that the audience doesn't choose one over the other but desires a fusion, appreciating his ability to apply a life philosophy to his business strategy. The AI identifies "life philosophy" as the driver for deep engagement and "business strategy" as the lever for credibility and sales, noting that business videos often have higher CTRs due to promised concrete results. This aligns with the presenter's concept of "flow" versus "library" videos. The presenter then asks about the adjectives his audience associates with his brand. The AI identifies: 1) "Authentic" or "True," particularly when discussing neurodivergence, appreciating his lack of pretense; 2) "Illuminating" or "Lucid," recognizing his analytical ability; 3) "Reassuring" or "Guilt-relieving," which he aims to foster; 4) "Inspiring"; and 5) "Strategic," noting his effectiveness despite an artistic side. His positioning is summarized as a guide to an "aligned life" with a core concept of "responsible freedom." He investigates his perceived strengths and weaknesses. Strengths include "radical lucidity," "synthetic mind," "vulnerable authenticity," and "minimalist aesthetic." Weaknesses identified are a "sense of repetition" in business themes, a "lack of technical proof" (showing the "behind-the-scenes"), and "perceived inconsistency" or being a "weather vane," though this is also seen by some as a charm. The AI also highlights frequently cited "click" phrases from his audience, such as "becoming a micro-celebrity," "I'm not lazy, I just function differently," and "work less to live better." Delving deeper, the presenter asks the AI to analyze comments for a specific video, "Why YouTube Can Change Your Life." The AI summarizes comments reflecting motivation for creators, nostalgia, loyalty ("Team Moustache"), pride in his journey, and concerns about shorts, fast cuts, and the survivorship bias. It also flags a technical bug he was unaware of. He then asks about the most watched passage in that video. The AI identifies "Nourishing the Algorithm" as the key moment, noting a peak in retention, suggesting viewers rewatched or were highly attentive to this strategic point about the algorithm for small channels. He then asks for improvements, and the AI points out a significant audience drop in the first 30 seconds, suggesting the intro is too long. It also notes comments perceiving his discourse as idealistic, recommending more raw discussion of failures. Analyzing the retention curves of his last 10 videos, the AI identifies that sharing "numbers and proofs," strategic explanations of the algorithm, and transitions to lists or new chapters maintain retention. Vulnerability, as seen in his ADHD video, also keeps retention high. Calls to action, particularly mid-video, and the concluding segments cause audience drops. The AI offers five tips to boost retention: curiosity loops, integrating CTAs into narration, pattern interrupts, reducing transition tunnels, and the "dividend effect" (one example per idea). The presenter explores using the AI for business strategy, specifically whether his YouTube videos bring in clients. He asks about audience drop-off during CTAs. The AI notes that in purely strategic videos, audience disengages when the topic shifts to external resources, contrasting this with longer, personal videos where CTAs are better tolerated. Offer announcements are better received on his ADHD video, and mentioning his newsletter at the end of longer videos causes a gentler drop. He learns that CTAs are best placed in the middle of videos, sandwiched between a tease and valuable information. He tests the AI's ability to extract testimonials from comments for a sales page, but the initial attempt with an ADHD video yields negative comments. He then tries an email video, getting a sarcastic comment about his newsletter. The AI identifies that the audience is intrigued by his newsletter ("les pépites"). For future video ideas, the AI suggests themes based on his top-performing videos: lifestyle and radical choices, psychology/mental health (ADHD), and YouTube strategy evolution for small creators. However, the presenter finds the suggested titles and topics too vague or too long, and prefers generating outlines himself. He probes for subjects of disagreement or potential controversy. The AI highlights the "paradox of the unstable trainer," where some question his credibility due to his frequent life/business strategy changes. Another point of contention is his portrayal of ADHD and self-diagnosis, with some feeling he makes too many shortcuts. His video critiquing the coaching industry also sparked debate. The AI notes that his perceived instability is both a source of frustration for some and loyalty for others, who see him as a "living laboratory." The presenter concludes that the AI is best used for brainstorming, understanding audience debates, sensitivities, and what generates clicks and attention. He believes creators should enter the arena with bold ideas rather than trying to appeal to a non-existent average avatar. He also suggests using the AI for creating evergreen tutorial videos based on recurring beginner questions. He tests the AI's ability to update old, high-performing business videos. However, the AI's context window is too short, limiting its analysis to recent videos rather than the older, highly successful ones he needs. It does, however, suggest updated angles for some business videos like "How to Create a Sales Funnel in 3 Hours." The presenter also asks about aspects of his daily life that interest people for vlogs. The AI identifies his "extreme minimalism challenge," his detachment from material possessions (referencing a video about moving), and his "TDH productivity vlog" as potentially engaging. It also notes audience interest in his moments of "bugs," strategies for focus, and use of his agenda. His radical disconnection challenges are also highlighted as divisive. The AI can also filter comments, ignoring praise or unconstructive criticism and providing actionable improvement points or identifying ideas for exploitation. It can also anticipate flops by identifying potential negative reactions to a video topic. Furthermore, it can check script clarity and accessibility, comparing technical terms and jargon against retention data. The presenter acknowledges the AI's vast database and useful features, but points out significant limitations: the small context window for historical data, the lack of macro-level trends across all channels, inability to process images, and a tendency to optimize for YouTube's metrics (views, watch time) rather than the creator's business goals (clients, revenue). He stresses that the AI is a supplementary tool and does not replace human coaching for conversion, sales, and personalized feedback. He promotes his own "YouTube Dividend" coaching program, emphasizing the importance of consistency and building a client-generating machine through YouTube.