Start with the English gap
Yinghanni works in precision manufacturing and needed a more dependable English content process. An export effort still required an English website, product content and outward-facing video.
I first worked on the website and purchasing information, then used HeyGen to create lip-synced English video. I connected video processing and publishing steps through n8n to reduce repeated manual work.
This was before Codex and Claude
I did not yet use my current AI development tools. Working through English content, video and automated publishing was part of how I gradually learned where AI could fit into a real company workflow.
The sequence matters: I did not learn a collection of tools and then look for a use. I encountered an operating gap and selected tools around it.
Automation still needed internal decisions
Tools cannot invent correct product facts. The company still had to confirm what the product was, which parameters could be public and what a buyer needed to know.
I connected the export content direction, website delivery, video creation and publishing workflow. Tools handled repeatable actions; people remained responsible for product truth and final judgment.
Confirmed operating results
The website and English content workflow went live. According to a retained Search Console review dated 26 August 2026, Google had indexed 22 canonical pages of the website at that time.
That figure is a dated search snapshot. It does not establish durable ranking, customers or orders.
What the project taught me
Enterprise AI does not always begin with a large platform. A clear operational gap—no repeatable way to produce English content—can be a practical entry point. Get the workflow running first, then decide what deserves further automation.
Visit the Yinghanni website or read the website release workflow.