Building AI-generated 'influencer' Characters for Affiliate Marketing, Using Psychology-driven Audience Research to Design Who the Character Is

Eric Swanson · How I Print $19K/Mo (Profit) With AI Influencers at 27 (Feels Illegal) ↗

Before designing any AI character, use Claude to research a specific, expensive, long-term problem affecting a large group of people (going beyond the surface symptom to the psychological pain and daily-life impact), then research who that audience actually trusts as an authority, and use a detailed JSON-style AI image prompt to generate a character matching that trusted archetype; post short-form content as that AI character addressing the audience's pain point, driving viewers to an affiliate link in bio or via a DM keyword sequence, and track commissions through the affiliate network's own order-and-click reporting.

The presenter says he made $19,000 profit in one recent month from AI-generated 'influencers' who don't exist, but is candid that his first two months using a generically attractive character produced only two sales and about 120 followers; the pitch is built around his own realization that character design must follow audience-trust research, not personal creative preference.

Someone comfortable directing AI research and image-generation tools iteratively and willing to do real audience psychology research before building a character.