Yazmyne Ortega / case study
AME Networks: automated marketing for insurance agents
Co-founder, technical lead (integrations and automation) · Sep 2026 to present · Pre-launch, launching Nov 2026
Make.comClaude APICloudflare WorkersCloudflare KVBrowser RenderingDNSEmail RoutingGoHighLevelGoogle SheetsPexels APIStripe
Overview
A subscription marketing service for independent insurance agents in network-marketing companies. Each agent gets a personal website, a branded social presence, scheduled social posts, email marketing, and a CRM, all generated and run automatically. The design goal was zero manual production: the two founders only approve, and the system does the rest.
The problem
Independent agents need a professional online presence to build trust and capture leads. Most don't have the time, budget, or skills to build one, and template-based services make every agent on a team look identical.
My role
- Co-founded the business in a 50/50 partnership. I own the technical side: automation, AI, hosting, DNS, payments integration, and workflow logic.
- Designed the end-to-end system architecture and an automation map covering 13+ workflows, from onboarding through billing, content, and monitoring.
- Set up business operations: EIN, entity structure, partnership agreement review, and a tiered pricing and referral model.
What I built
- AI website generator. Takes an agent's onboarding answers and produces a complete, unique, mobile-ready website, published live to their own subdomain within minutes. No templates and no manual design.
- Uniqueness engine. Each site gets its own layout, palette, fonts, section order, and writing voice. The generator checks every live site's headings and photos and steers away from them, so agents on the same team never look or sound alike.
- Copy quality controls. Prompt engineering and post-processing rules that remove AI-sounding and banned phrasing, so the copy reads the way people actually talk.
- AI social post writer. Monthly posts with similarity checks, an automatic rewrite pass, and stock photo matching.
- Facebook cover generator. Designs a cover that matches each agent's site and renders it to an image automatically.
- Lead relay. Website form submissions flow straight into the agent's CRM and pipeline.
- Serverless hosting. One edge worker serves every client site under a wildcard subdomain. Publishing or taking a site offline is a single API call.
- Client approval flow (designed). Clients approve or regenerate posts asynchronously, with auto-approval at the deadline, so there are no meetings.
Results so far
- 5 demo agent sites generated and live, with 0 shared photos and every site loading successfully.
- Per-client AI cost of cents for setup and about $0.10 a month to run.
- Platform infrastructure runs on roughly $16 a month plus usage-based AI, designed to scale past 100 clients without a plan change.
- Iterated the generator through 7 versions based on test-round findings, systematically eliminating repeated wording across sites.
See it
tiffany-nguyen.amenetworks.com. A demo site for a fictional agent, generated end to end by the system.
← back to portfolio