The video opens with a bold claim — people quietly making thousands with Claude-built IA influencers, $100k in 30 days, $2k on a single sale, $200k profit — but is it real or just IA hype? Mark Tilbury, a self-described old-school millionaire, gives himself seven days to document everything and share raw numbers, promising to pay $1,000 from his own pocket if he fails to clear that bar. Three constraints keep it honest: under $99 to launch, no face or voice of his own, and the first IA influencer must be ready before the day ends.
First move is competitor reconnaissance. The most famous case is Lil Miquela — over 2 million followers, Met Gala and Grammys appearances, brand deals with Prada and Calvin Klein reportedly driving about $11 million a year, positioned as a full supermodel. Tilbury logs that under brand deals and notes the ceiling: this format can be taken extremely far.
Next is Mother Satori — presented as a 108-year-old U.S. holistic guide. Health is framed as a perennial money niche, and her loop is simple: daily videos ending with a single gut-cleanse drink, thousands of views a day, comments full of purchase asks. Even one hero product at that attention level implies meaningful revenue.
Du faux menuisier à la mamie IA : un marché saturé
A quick sweep widens the map: a glam lifestyle creator doing relationship podcast reels, a patriotic British woodworker handcrafting flanks who is obviously not real, and an IA granny doing outrageous stunts with over 2 million followers. Tilbury flags the ethics — deceiving customers is not ok — but the monetization breadth is undeniable, from brand deals to product pushes.
With that landscape, Tilbury picks niches. The three money verticals he trusts are health, wealth and relationships; two are already proven on the board. Curiously, none of the IA accounts he found talk about money, so wealth looks like a gap. To hedge, he creates one influencer per niche: Amos for health, Vivian for wealth, Sienna for relationships.
He outsources character ideation to Claude on the free plan: 'generate a health-niche synthetic influencer profile — name, backstory, distinct look, unique selling point.' Amos appears — a mysterious 101-year-old, shaved head, long beard, heirloom remedies, a male mirror of Mother Satori. Wealth gives Grandma Vivian — started from nothing, three husbands, private jets, head-to-toe designer, a bougie boss grandma opposite every rented-Lambo young man in the niche. Relationships gives Sienna — 22, stylish, talking dating and situationship drama, close to the earlier Olivia archetype. Tilbury remains skeptical but the roster is set.
Higgsfield rencontre Claude : du fond blanc au porche
For visuals, Higgsfield is chosen after a short tool comparison. Sign-up via email, questions answered as personal use, goal create avatars, experience intermediate. A pop-up offer is skipped because Tilbury already negotiated a discount code TILBURY for his audience, claimed under profile → manage account → promo code. Among Starter, Plus and Ultra, Ultra wins because it offers over a hundred video generations and still fits the $99 budget.
Claude and Higgsfield are then wired: Claude → customize → connectors → add custom connector, paste the link, and the two chats talk. For Amos, Claude writes a monster prompt covering multiple angles, full-body and close-up, white background, realistic skin rather than airbrush — the first render is too young, so 'make him a few years older' yields a better fit. Environmental plates follow: porch, farmhouse kitchen, vegetable garden and yard, all plausibly Amos. The loop repeats for Vivian — fur coat, sunglasses, jewelry, likened to the 101 Dalmatians villain, jet, sports car, drawing room and terrace for tea and scones — and for Sienna — stylish and convincingly real, with matching domestic locations.
Mapping viral content without hand-writing scripts, Tilbury leans on a Claude skill called content engine, installed via customize → add → upload skill and invoked with '/'. Asking it to generate content ideas for Sienna returns a full list without manual ideation; he likes idea 2 for relatability, idea 1 for advice, idea 7 for a psychological reveal, all suited to podcast settings. He generates three scripts for Sienna, then repeats the process for Amos and Vivian, ticking scripts off the list.
Voices are built in Higgsfield's audio tab: upload Vivian images, type a short direction like a short Vivian-style voice instruction, and ask Claude to draft a detailed voice prompt to paste into Higgsfield. The generated voice lands as an old-school teacher tone — good enough. The same pipeline clones voices for Amos and Sienna.
Bringing images, locations and scripts to life happens in cinema studio → new project named Sienna, aspect 9:16 vertical, quality 1080p. A first non-talking B-roll is generated from a detailed prompt referencing image one, the Sienna character sheet, to test consistency for later text overlays. Additional B-roll and talking clips follow; Tilbury previews favorites: 'early to bed early to rise' over coffee steam, 'the number one longevity food is beans, cornerstone of every long-lived culture' despite the wife's doubts, 'I can't turn water into wine but I can turn you into mine', 'she was using you' on friendship betrayal, and Vivian's timeless wealth line — 'buy the ugliest property on the nicest street, fix it properly, rent or sell, roll the proceeds, repeat for 40 years, don't touch it, let it buy the next one.'
Exhausted but on schedule, Tilbury notes the first influencer was built within a day. Next morning, captions are added with free tools like CapCut or Instagram Edits, and accounts launch on Instagram only, deferring TikTok and YouTube Shorts. For Amos he tries Amos → health and wellness, username taken so Amos.oldways because he preaches the old ways, picks a lamb photo as avatar, and writes the bio as 101 years old, simple habits for peak energy, focus and longevity, explicitly adding 'your IA grandpa' for transparency. He repeats the setup for Sienna and Vivian, posts ten videos per account, roots for Vivian, and waits a couple of days to crown a winner. An early comment thread is revealing: one user bashes Sienna for a Shein dress and a $3 frappuccino, another defends her, both arguing without realizing she is not real, while others ask where she bought her clothes — the realism passes the casual test.
A few days later the leaderboard is clear: Vivian third — a sting for Tilbury but the reason he launched three — Sienna second with genuinely impressive numbers for a new account, Amos first by a gap. He crowns Amos and reviews monetization paths noted at the start: brand deals (out for this sprint due to outreach lag), physical products and Shopify (a project in itself), ad revenue (not available for Instagram reels), affiliate via TikTok Shop (no TikTok presence, hindsight error), leaving digital products as the fit for an educational figure like Amos. He rejects generic ebooks because the same basics are free via Google or ChatGPT; the product must be structured, interactive and transformation-oriented. Back in Claude he invokes the digital product skill, asks for a workbook for Amos, gets three concepts, picks the 30-day challenge with one habit a day, and builds it — day-by-day rules with tick boxes, a fridge tracker, and bonus mother's dinner recipes. He adds an Amos photo to the cover; a character that did not exist five days ago now headlines his own book, which he vets for safety and removes any IA slop.
Selling is wired through Stand Store on its 14-day free trial, framed as non-sponsored research: My Store → add products → digital products → style call out, thumbnail using the sheep image matching the Instagram avatar, title 'get my 30-day energy reset', button 'start the old ways challenge', price $15.99 discounted to $12, upload, publish, copy link to Instagram bio, and connect payout off-camera. On the final day he also surfaces a big lesson: after doubling down on Amos, Vivian exploded — over 10k views on one video, 20k on another, and a 250k-view viral hit. He admits he should have trusted his gut and notes a seven-day sprint is great for testing but sometimes needs more time for the full picture; still, every influencer gained traction and not a single comment complained about IA use, which surprises him given his own reservations about IA slop. He is explicit he does not want everyone to jump on this because social media's soul is real human connection, and IA everywhere would hollow it. Numbers: 23 sales at $12 equals $276 revenue; costs were Claude free (with waiting for credits), Instagram Edits free, Stand Store free trial zero, PayPal fees $19.52, Higgsfield at full Ultra price $99 for fairness despite the Tilbury discount, total costs $118.52 and net profit $157.48. He frames $157 as not life-changing but telling for minimal effort and budget, and sketches the upside — reposting the same library to TikTok and YouTube for incremental reach without extra shoot, and trying the other monetizations across all three influencers. A deal is a deal, so he will pay the promised $1,000 to a commenter who names their favorite influencer and why, with no Vivian bias, and teases a seven-day IA dropshipping attempt next.
Commentaire de l’IA
"À mon avis, la vidéo démonte le mythe du 'paresseux' : les outils sont paresseux, pas les choix. Repérer le créneau vide en wealth, couvrir le risque avec trois personnages et miser sur un seul workbook est la vraie stratégie, et 157 $ prouvent que la boucle fonctionne."
Évaluation de l’IA
À mon avis, la force de la vidéo est de tester 'paresseux' comme hypothèse : elle valorise les décisions — repérer le vide en wealth, différencier les personnages, miser sur un workbook — plus que l'outil. La victoire d'Amos le confirme. Mais la fenêtre de sept jours et dix vidéos par compte comprime la découverte — l'explosion tardive de Vivian à 250k montre le prix d'un sacre précoce.
Les limites sont claires : mentionner 'your AI grandpa' aide la transparence, mais la plupart des abonnés n'ont rien vu, et des conseils santé sont vendus sans discussion sur l'exactitude médicale ou les avertissements. Le calcul des coûts est idéalisé — attente des crédits Claude, Higgsfield compté plein tarif, sans taxes ni temps.
Pour défendre l'autre camp : les sceptiques du 'argent facile' ont raison — 157 $ ne remplacent pas un salaire et sans tester TikTok/Shorts, marques ou affiliation, généraliser est prématuré. Ma conclusion : comme test marché à petit budget, la boucle est viable ; l'échelle vient de la répétition et de la réputation.
Sources
6 liens ; aucun autre article publié ne les cite. Stories sharing a link do not confirm each other; a source's origin is not inferred from how often it is cited.
influenceur ia · higgsfield · claude · instagram · produit numérique · side hustle · nodesdaily