Use cases
Four real ways to use the AfiniTwin daily. Composite characters. Any resemblance to real persons is coincidental. (Closed decision 2026-05-10: public cases with marked fictional characters, no real testimonials in Phase 1.)
Ana — the direct consultant
Mi Afini Standard in Claude Projects
- Ana downloads the .md of the Standard preset in Spanish.
- Creates a Claude Project named "Mi Afini Standard" and pastes the Project instructions.
- Uses it for daily conversations: planning, drafting, reviewing client proposals.
- Activates Voice mode when she needs sign-able drafts ("write this for me").
- Activates Analyst mode when a client asks about psychometric validity of a test their company uses ("give me the technical analysis of my Big Five").
- After the second quarter, repurchases Twin with updated data.
Pablo — the ghostwriter of himself
Mi Afini Ghostwriter in ChatGPT Custom GPT
- Pablo downloads the .txt of the Ghostwriter preset.
- Creates a Custom GPT called "Me, writing".
- Pastes the content in Instructions.
- Uses it to reply to long professional emails: Pablo provides the idea, the GPT turns it into sign-able text that sounds like him.
- Doesn't use it for intimate messages (he writes those).
- If the GPT inserts a verbal tic Pablo does not use, he gives feedback: "no, that word is not mine".
Marta — the reflective decision-maker
Mi Afini Critic in Gemini
- Marta downloads the .md of the Critic preset after reading and accepting Appendix B.3 disclaimer.
- Creates a Gem called "The other side".
- Activates it when she has an important professional or personal decision on her mind.
- She pastes the argument in favor of the decision and asks for critic mode.
- The Gem returns the best possible refutation. If Marta still thinks the same after, she's more confident. If not, she delays the decision.
- Does NOT use it when tired or emotionally loaded — Appendix B.3 advises against.
Hugo — the technical user
Full JSON via Anthropic API
- Hugo is a developer. Downloads the full JSON with opt-in narratives only for professional context.
- Loads it in his personal agent via Anthropic's direct API.
- Leverages prompt caching: the system prompt is cached across turns, cheapening cost.
- Combines with his own tools to make the agent help him code while keeping his technical voice.
- When he repurchases Twin, he runs a git diff between the two JSONs to see how his profile changed.