LensAI Engine

Inspiration

What it does

How we built itInspiration

The optical retail industry suffers from a massive bottleneck: the friction of the sales process. Customers receive complex, hard-to-read optometry prescriptions and drop out of the sales funnel because quoting lenses and trying on frames requires physical visits or slow human intervention. By analyzing user behavior and the psychology of digital sales, we realized that instant, frictionless responses build the highest trust. We wanted to eliminate this barrier by turning a simple photo of a prescription into an immediate, personalized sales experience right where the user already is: WhatsApp.

What it does It is an autonomous AI sales agent and quoting engine for optical clinics. The flow is seamless:

The customer sends a photo of their handwritten or printed optometric formula via WhatsApp.

Our system processes the image using Gemini's multimodal capabilities to extract exact clinical values (Sphere, Cylinder, Axis).

The AI instantly matches these values with the clinic's inventory, generating a precise quote and explaining the lens options in natural, persuasive language.

A seamless handoff to a React-based web interface for an AR "Try-On" experience, acting as a real-time virtual image consultant.

How we built it We orchestrated the core backend architecture using n8n to handle the conversational flows and webhooks from the WhatsApp Business API. The brain of the operation is the Gemini API, utilizing strict system instructions to guarantee structured JSON data extraction from clinical images. We are deploying this infrastructure under the umbrella of Creative Code Labs, ensuring an enterprise-grade, scalable environment that currently serves real B2B clients. The upcoming AR interface is being developed using React and Node.js.

Challenges we ran into Parsing highly variable, often blurry handwritten prescriptions from different doctors. We overcame this by fine-tuning our prompts and utilizing Gemini's advanced multimodal reasoning to contextualize the clinical data accurately before outputting the structured payload.

Accomplishments that we're proud of Securing real-world traction before even finishing the competition. This isn't a sandbox prototype; the architecture is designed to be immediately deployed to two active optical clinics, guaranteeing real users, real financial transactions, and measurable business impact from day one.

What's next Scaling the AR Try-On feature and transitioning the system from a single-tenant solution to a fully scalable SaaS model, allowing any small optical business to onboard their inventory and deploy their own AI sales agent within minutes.