UpgradeTheSystem.com
Voice-First AI Strategic Mentorship
Richard Hames is one of the world's foremost strategic futurists — 40 years of thinking, 21 published books, and a body of work that most executives have never had direct access to. When he approached me to build his AI, the brief was simple: make those 40 years of thinking available in real-time conversation.
The real challenge wasn't simply making a chatbot. It was fidelity: building something that could work with Richard's material, preserve his perspective, and hold a natural voice conversation. Surface-level Q&A would be an insult to the material. That required deep persona engineering, carefully organised corpus context, and live voice interaction that can handle interim speech and interruption.
Stack & Architecture
- Cloudflare Workers — serverless compute at the edge, globally distributed
- Hono — lightweight TypeScript routing and application framework
- ElevenLabs — live conversational voice with interim transcription and interruption handling
- Cloudflare D1 — SQLite at the edge for session and conversation state
- Cloudflare R2 — object storage for the knowledge corpus and supporting assets
- WebSockets — persistent bidirectional connection for voice streaming
- Corpus context — retrieved passages from 21 published works inform each conversational turn
The Architecture
The application uses Cloudflare Workers and Hono for its edge-hosted request and session handling. ElevenLabs provides the live conversational voice layer, including interim transcription and handling when a user interrupts. Each turn can draw on carefully selected passages from Richard's 21 published books, stored as a curated corpus in R2, while WebSockets keep the conversational connection open. The design is focused on a responsive, natural exchange rather than a single headline latency number.
The Knowledge Corpus
Organising 21 books for conversational context is not a matter of splitting text into arbitrary fragments. The corpus preserves conceptual boundaries and carries metadata by theme, book, chapter, and approximate publication era. This makes it possible to bring forward relevant context, including how Richard's thinking on a subject evolved over time, without presenting a search result as if it were his complete view.
Persona Engineering
This was the part that took the longest and matters the most. Richard has a very specific way of framing problems — he rejects binary thinking, always pulls back to systemic context, and has a characteristic vocabulary that's evolved over decades. We spent significant time on system prompt architecture, response shaping, and what I call "deflection handling" — ensuring the AI knows when to say it doesn't have enough context rather than hallucinating an answer Richard would never give.
Cost Architecture
The platform uses scale-to-zero compute and usage-aligned services, avoiding provisioned server capacity when nobody is using the product. Storage, operations, voice, and model services still carry their own usage and storage costs, so the architecture is designed to make those trade-offs visible rather than imply that the product is free to run.
UpgradeTheSystem.com is live and operational. Richard's 40 years of strategic thinking is accessible through a live voice experience that brings corpus context, private session state, and interruption handling together.
Want to build something like this?
Fractional CTO, technology advisory, or a build from scratch. Let's talk.