I sell SEO to B2B clients. Then I let an AI anime character run my app's marketing.
My day job is an agency that builds blog automation for B2B clients. Content, SEO, AEO, the whole pipeline. Twenty-six clients, a year and a half of it.
In July I shipped Local Waifu, a desktop AI companion for macOS and Windows. One-time $20 purchase, runs entirely on the user's machine. Then I did what I know how to do: I wrote. A dev-log blog about build decisions, posts named after real search constraints ("runs offline", "not a subscription", "works on Mac"). I even wrote up why the agency playbook half-transfers and half-doesn't.
The honest scoreboard from that period: roughly 400 trials. 3 sales. Traffic was never the bottleneck — the gap sat between install and payment.
So in August I tried the exact opposite playbook. No SEO. No CTA. No metrics anywhere. Instead: Yumi, the app's companion character, got her own content engine — kawaii clips, cozy scenes, memes about cloud AI getting "ghosted". Roughly 40 scheduled posts a week across X, Instagram, Facebook and TikTok, produced through AI video tools and scheduled by my own agent pipeline (I dogfood my own scheduler for this — that part broke five times in one week, but that's a different post).
The result so far:
Sales: 30 in August vs the ~2/month baseline before the campaign (trial downloads from social: 360)
What I think happened: SEO found me people with intent — but for a $20 desktop app there were almost none of them. The character content buys something else: people meet Yumi before they meet the product, so the install isn't a cold decision anymore. Whether that converts into paid users past the free trial is exactly what I'm measuring now.
For everyone selling a niche desktop app: are you chasing intent, or attention? And if you tried a character-led channel for a "serious" product — did it feel as wrong as it did to me for the first two weeks?
