How Mindvalley Drove $19M+ in Annual Sales Through Web Architecture and Conversion Optimisation
A deep-dive into the web architecture, conversion optimisation and AI platform work behind the largest revenue-impact project in my career, and what it means for your AI rollout.
Driven through web architecture and conversion optimisation, including the Storyblok page builder (469 PRs)
Across 17 repositories: Next.js, Storyblok plugins and the Elixir/Phoenix backend
BlinkWork Turborepo monorepo: NestJS, Next.js, TypeScript, PostgreSQL, with developer mentoring
The Challenge
Mindvalley, a global edtech company serving millions of users across 120+ countries, needed web architecture and conversion optimisation that could keep pace with the business, from the Storyblok page builder to the Next.js frontend and the Elixir/Phoenix backend behind it.
The team also needed someone who could not just ship the web platform, but help build BlinkWork, Mindvalley's B2B multi-agent platform on MCP, with real-time AI chat and RAG-based memory.
My Role & Approach
Drive web architecture and conversion optimisation: 745+ PRs merged across 17 repositories, spanning Next.js, Storyblok plugins and the Elixir/Phoenix backend, including the Storyblok page builder (469 PRs) behind $19M+ in annual sales.
Build BlinkWork's connector framework: a two-way connector framework for 20+ integrations, from BigQuery and Slack to Cloudinary, with vault-stored credentials, in the Turborepo monorepo (NestJS, Next.js, TypeScript, PostgreSQL).
Ship real-time AI chat and organisational memory: SSE streaming controller, system prompt builder, tool invoker and conversation persistence, plus a memory layer on TrueMemory RAG and knowledge graph APIs with search and ingest MCP tools.
Introduce the Jev AI model: brought the Jev AI model into BlinkWork and demoed it to the whole tech team, while mentoring developers across the monorepo.
Technical Architecture
The 5-Layer Platform Model (applied at Mindvalley)
Content feeds the storefront, chat reaches memory, and both sit on the same services and data.
01 · Content
Storyblok headless CMS and page builder for conversion-optimised pages
02 · Frontend
Next.js, TypeScript, Tailwind CSS, and Magic UI
03 · AI
SSE chat with a prompt builder, tool invoker, and the Jev model
04 · Memory
TrueMemory RAG and knowledge graph, with MCP search and ingest
05 · Backend
NestJS and Elixir/Phoenix on PostgreSQL, with Docker and GitHub Actions
Full Tech Stack
Frontend
AI & Backend
Infrastructure
Key Outcomes
Drove $19M+ in annual sales through web architecture and conversion optimisation, including the Storyblok page builder (469 PRs).
Shipped 160+ PRs in the BlinkWork Turborepo monorepo (NestJS, Next.js, TypeScript, PostgreSQL) with developer mentoring.
Built a two-way connector framework for 20+ integrations, from BigQuery and Slack to Cloudinary, with vault-stored credentials.
Built real-time AI chat from scratch: SSE streaming controller, system prompt builder, tool invoker, and conversation persistence.
Built an organisational memory layer on the RAG and knowledge graph APIs of TrueMemory (Mindvalley's memory engine) with search and ingest MCP tools for agent retrieval.
Introduced and demoed the Jev AI model, used in BlinkWork, to the whole tech team.
Want similar results for your platform?
Book a 30-minute strategy call and I'll show you how web architecture and conversion optimisation decisions translate to revenue outcomes, and where AI fits in, specific to your context.
Also available: AI Readiness Scorecard · CHRO People-AI Checklist
Director-level references from Mindvalley: I am actively reaching out to former Directors and VPs for business-outcome testimonials. If you'd like to speak directly with a Mindvalley engineering leader, mention it in your call booking and I will arrange an introduction.