Fashion.at’s AI audience twin sketches subscribers, but cannot predict their response

Cartoon · AI-generated
A Gemini-powered simulation proposed audience profiles and subscription tactics for the magazine, but its fictional persona and campaign recommendations have not been tested with real readers.
Fashion.at has used an AI-based digital-twin tool to construct a hypothetical portrait of its readership and simulate how that audience might respond to an annual subscription offer. The exercise produced a fictional figure, Elena B., a 36-year-old Vienna-based curator, intended to represent a broader group aged 28 to 48 rather than any identifiable subscriber. Fashion.at says the tool was developed by Dr Josef Sawetz and tested using Google’s Gemini 4 Argon model. Google describes Argon as designed for complex, extended tasks, but its product information does not verify the magazine’s experiment or its conclusions.
The imagined audience, labelled “The Culture & Design Cosmopolitan”, is characterised by an interest in Austrian fashion, contemporary art, cultural history and sustainability. Its profile also includes scores across the Big Five personality dimensions, with openness to experience and conscientiousness rated particularly highly. The framework is a way of grouping broad personality traits, as Oxford Reference explains; the scores themselves are generated estimates, not psychological assessments of real readers.
Fashion.at then asked the tool to assess a fictional annual subscription campaign. In one version, subscribers received an imagined bag designed for the magazine. The model judged the gift consistent with the publication’s cultural and design focus, rather than treating it simply as a sales incentive. In a second scenario, the bag was removed and the analysis extended to audiences beyond Austria. The magazine says the tool recommended foregrounding exhibitions, cultural events and its archive for readers in Vienna and the DACH region, while positioning the English-language offer internationally around independent European design journalism and curated cultural perspectives.
The analysis also identified the United States, Germany, India and East Asia, particularly China, as markets of interest. For the campaign itself, the model favoured avoiding steep discounts and urgency-driven prompts, and suggested a two-month benefit with annual billing. These remain recommendations from a simulation: Fashion.at says it has not tested the messages in a live campaign, and the bag exists only in the fictional scenario. Google’s description of Gemini 4 Argon establishes the model’s intended capabilities, not the reliability of these particular audience predictions.
The experiment draws on research into psychological targeting and motivational message matching, but that scholarship does not establish that Fashion.at’s persona reflects its actual readership or that its proposed subscription messages will succeed. A study by Sandra Matz and colleagues, published in PNAS, found that matching some advertising appeals to psychological traits could affect clicks and purchases in other settings. A 2022 review by Joyal-Desmarais and colleagues likewise found an average positive effect from matching persuasive messages to people’s motivations, while noting variation across studies and contexts. Fashion.at’s exercise is best read as a model-generated interpretation of its editorial identity, not evidence of reader behaviour: the magazine’s imagined audience is addressed as culturally curious, with fashion treated as a route into art, history and society.
Inspired by headline at: [1]
Sources by paragraph: - Paragraph 1: [2], [3] - Paragraph 2: [2], [6] - Paragraph 3: [2] - Paragraph 4: [2], [3] - Paragraph 5: [2], [4], [5]
Source: Noah Wire Services · Original