Juan contracted with an optometry company that needed to reduce staff input errors. The work became a Chrome extension aimed at forms on their vendors' sites, not a generic everyday-form tool.
Staff entered the same kinds of values into vendor workflows under time pressure, and small mistakes created hours of cleanup later. The extension sat in the browser and ran rule-based AI against the fields that mattered. It could trigger warnings when something looked off, prevent destructive actions before they landed, and alert managers when a pattern needed attention.
It did not claim to eliminate every mistake. It made the costly ones harder to complete unnoticed.
What it focused on
- Watching vendor-site forms where bad input hit downstream systems
- Warning in context when values were missing, duplicated, or outside expected patterns
- Blocking destructive actions that would be hard to undo
- Alerting managers when staff activity crossed a threshold that needed review
- Keeping logic local to the browser so the team could adopt it without a heavy install project
Contributions
Juan implemented several new features, including a rich text editor, worked on the management backend, and optimized Chrome extension performance.