Anne: privacy-first AI that earned 1M+ downloads

I own Anne end-to-end — a privacy-first AI assistant built for an underserved community. Consumer AI is crowded; Anne's wedge is trust and cultural fit, not generic chatbot capabilities. My mandate covers product wedge, growth efficiency, and the retention story in the data room. Downloads are vanity if retention and CAC don't hold.

Mainstream AI assistants treat privacy as a settings page and community as a marketing line. Users who need culturally aware, privacy-respecting AI had no product that felt built for them — only adapted from products built for everyone else. The trap was chasing feature parity with general assistants. We'd lose on breadth and win nowhere on trust.

I bet on privacy-first architecture and community-specific design as the product wedge — not feature parity with general assistants. Retention and acquisition efficiency would prove the wedge; download count alone wouldn't. Every data collection decision went through: does this earn trust or spend it?

[NEEDS YOUR INPUT: Add the specific trade-off — e.g., what data you chose not to collect, or a feature cut that preserved privacy/retention.] Until then: the fight was resisting general-purpose AI feature creep that would have diluted Anne's trust positioning. Growth teams wanted viral loops; I wanted retention loops.

Anne reached 1M+ downloads with low CAC and strong retention. I owned the growth and retention story for the seed-stage data room — metrics, narrative, and proof that community-specific AI can acquire and retain at consumer scale. Trust showed up in the numbers, not just the positioning doc.

I'd publish a plain-language privacy promise users can audit, not legalese they ignore. I'd gate viral features behind retention milestones — novelty spikes downloads; trust keeps them.