




Tools & Technologies Used

Frontend
Next.js

Content Management
Sanity CMS

Project Management
Jira

SEO Monitoring
Google Search Console

Analytics & Tracking
Google Tag Manager

Hosting & Deployment
Vercel

UI/UX Design
Figma

Frontend
Next.js

Content Management
Sanity CMS

Project Management
Jira

SEO Monitoring
Google Search Console

Analytics & Tracking
Google Tag Manager

Hosting & Deployment
Vercel

UI/UX Design
Figma

Heyou Features

Challenges Faced:
- LinkedIn does not want to be automated: Pass Chrome Web Store review, run cleanly across team accounts, and earn a strong store rating from day one.
- Make engagement look human: Use AI automation to replicate natural behavior patterns so LinkedIn's algorithm cannot distinguish Heyou's activity from a real user's.
- Protect user accounts at scale: Catch account safety regressions before they reach production, across every post type and team configuration Heyou supports.
Our Approach:
- Chrome extension on Manifest V3: We built the extension from scratch on MV3. It detects new LinkedIn posts from configured profiles and company pages, queues engagement actions, and fires them through content scripts using the same DOM event pathways a real user click produces. No LinkedIn session cookies are accessed. Nothing leaves the user's browser that should not.
- AI agentic automation: Fixed intervals get accounts flagged. Our AI automation layer models each user's natural engagement cadence based on their account's own activity history, then varies timing, sequencing, and volume so the pattern looks like a person, because it is shaped like one. It adjusts dynamically as post volume and account behavior change over time.
- AI QA Testing: Our Automated QA identifies LinkedIn UI elements by function and context rather than brittle CSS selectors. When LinkedIn ships a frontend update, the system adapts rather than failing silently. Every release runs through automated scenarios across multiple account types, post categories, and engagement volumes before it reaches production.
- Chrome Web Store launch: We managed the full submission process: permission justifications, privacy policy alignment, and store listing copy. The extension launched with a 4.7 star rating and has held a clean violation record through every update since.
What We Delivered:
- Production MV3 extension: A fully published, compliant extension handling auto-engagement across company posts, personal profiles, and promoted LinkedIn content.
- Agentic engagement engine: Per-account AI automation that models realistic human engagement patterns and adjusts timing and volume to stay inside safe thresholds.
- AI driven QA framework: Adaptive QA that identifies LinkedIn UI elements by function, maintaining test coverage even when LinkedIn's frontend changes.
- Chrome Web Store launch: Full submission management, permission documentation, and listing work resulting in a clean launch and strong initial store rating.
Heyou Team Review:
What impressed me was not just what they built, it was how fast they understood the problem. LinkedIn automation is not a standard engineering challenge. They got that on day one, and the extension they shipped reflects it. We went from zero to a 4.7 star rating on the Web Store without a single revision request from Google.