Overseas Growth · 中文版 繁 / 简

Zero followers to
51,000 monthly reach in North America

I'm Vernessa. I currently work remotely on overseas growth for an AI companion startup, running English content on X and Threads and talking with users in their communities. Before that I owned the North American market for PicWiz, an AI photo editor, from positioning and content strategy to making the videos myself. Below is the whole story, including the stretch where nothing worked.

vernessa0607@gmail.com Based in Taipei · working remotely, remote first English and Mandarin · TOEIC 915
51,000Best monthly reach
33,100Top single video views
60+Videos in the cold start
4Videos over 10K views
Now

Overseas growth at an AI companion startup

Remote · Sep 2026 to now · X / Threads / Reddit · English-speaking users

What I do

  • Run an English lifestyle account every day. Daily posts on X and Threads, mostly about everyday life, with the product showing up on one fixed day a week. A weekly theme for each day (like a Saturday travel series) keeps me from running dry and tells followers what to expect.
  • Talk with real people in their communities. Every day I find English threads about AI companions and reply or send DMs. I check each profile first, stay away from minors and sensitive topics, and don't drop ads into other people's threads.
  • Play by each community's rules. On Reddit I read the rules first and join discussions for real. No vote trading, no accounts boosting each other.
  • Make short promo videos. An English video series built from script and character design to AI-generated reference images and clips, put together five seconds at a time.
  • Pitch features with something you can click. I noticed a daily-reflection habit in how people use the app, built a clickable phone demo of the feature, and pitched it with that instead of a doc.
  • Log everything. What I posted, who I replied to and which angle I used all go into the team's tracking sheet, so we can see which topics actually work.

How I judge whether it works

Not by likes. I care more about which topics bring replies and DMs, because a companion app needs people who want to talk, not people who tap a heart and scroll on.

Case Study

PicWiz cold start in North America

Xuantu AI · Overseas operations · Mar to Jun 2026 · YouTube / TikTok / Instagram, starting from zero followers

The problem

The product's messaging was written for Asian markets. Translated straight into English, it talked about features Western users didn't care about, in a tone that didn't land. The accounts started from zero, with no budget and no existing audience.

What I did

  • Decided again what we were selling. Instead of reusing the old marketing language, I looked at which editing needs Western users actually had and picked features and angles from there.
  • Turned "post and hope" into a weekly decision framework. I broke down user behavior on TikTok, Instagram and YouTube and turned it into a weekly plan: which topics, which hooks, which platform.
  • Ran the whole line myself. Strategy, topics, scripts, filming, editing, publishing and reading the data, with no handoffs.

Results

  • 5,500+ views a week in the first month
  • Best month: 51,000 reach
  • 60+ short videos, 4 over 10K views, top video at 33,100
The part that matters most

For a while, nothing I posted got views

It didn't slowly decline. It just stopped. The easy move is to keep posting and hope the numbers come back. I didn't do that.

1 · Separate the causesSplit it into content, timing and audience instead of blaming luck.
2 · Read my own dataCompare the videos that worked with the ones that didn't. Was the gap in the first three seconds or the topic?
3 · Study othersLook at similar accounts that were growing: how they hook and how they pace.
4 · Change one thing at a timePost again and confirm which change brought views back, so it isn't a lucky guess.

Views came back. I think this says more about how I work than the numbers do: I don't panic when the data is bad, I break the problem apart, and I'm willing to admit my first approach was wrong.

Accounts you can check

The three accounts I ran for PicWiz. The content is still up.

YouTube @bananaintern-q1g TikTok @internbanana Instagram @bananaintern

What makes me different

Lots of people make content. Fewer growth people build their own tools.

  • I build products. Over the past year I built 20+ AI apps on my own with AI coding tools, most of them deployed and usable. It helps me know what's quick and what's slow when I talk with product and engineering.
  • AI tools are how I work, not a bonus skill. Content production, data cleanup and workflow automation, I handle them myself.
  • I pitch ideas, then build them. When I have a feature idea, I bring a clickable demo. The conversation goes much faster.
  • Biomedical engineering background. I analyzed 50+ competing medical needle products, so I'm used to mapping a market before acting.
  • English and Mandarin. TOEIC 915. Everything for the North American market was written in English.