I built an AI that turns happy customers into Google reviews (real ones).
Hey everyone ๐ Long-time learner here. Recently I built a small but surprisingly powerful AI system, and I wanted to share the thinking behind it with this group. Most local businesses donโt have a traffic problem. They have a trust problem. 93% of customers read reviews before buying โ but only ~5โ10% of happy customers ever leave one. So the review profile most businesses have isโฆ skewed. Loud minorities. Silent majority. Instead of trying to โincentivizeโ reviews (which breaks trust), I built a mobile-first AI review generator that: โข Captures feedback while the customer is still on-site โข Helps customers turn their own words into a well-written review โข Lets the customer fully edit + submit it themselves โข Stays compliant with Googleโs policies No fake reviews. No automation posting on behalf of users. Just friction removal. From a systems point of view, this was interesting because: โข Itโs serverless (Cloud Run โ scales from 0 to spikes) โข Uses LangChain purely as orchestration (not over-engineered) โข Gemini 2.0 Flash tuned for human tone, not robotic outputs โข White-labelable across niches (dentists, salons, clinics, etc.) Early results show 5โ10ร increases in review volume, which then compounds into: Better Maps rankings โ more customers โ more reviews. I recorded a full breakdown + live demo here if anyoneโs interested No pitch. Just sharing the system and architecture in case it sparks ideas for: โข AI agency offers โข Local business automations โข Productized AI services Would love feedback from builders here: If you were selling this, would you position it as a product or a service? Curious how others in this community would package it.