We started building WIZWIKI before anyone formally assigned it to us. After joining a marketing company, we kept hearing the same complaints about SaaS CRM platforms: knowledge was scattered, automation was hard to inspect, and every useful feature seemed to require another subscription. We saw the need for something larger than a chatbot or integration—a marketing system the company could own, teach, and operate itself.

WIZWIKI brings CRM records, calls, customer conversations, campaign history, and approved documents into one system. Its DOJO runs complete simulated conversations and evaluates them for accuracy, tone, compliance, retrieval quality, and business value. Codex helps trace failures across prompts, knowledge, scoring, code, and tests, while people remain responsible for what the system learns and what it is allowed to do.

The architecture can distribute AI work across networked company computers rather than depending entirely on hosted services. Alice Brain turns Mac and Linux machines into outbound-only local AI workers, while Rotary Relay coordinates their availability, capabilities, work claims, and completion receipts. Workers receive narrowly scoped jobs without receiving production database credentials or infrastructure keys.

Alice Brain: https://github.com/curiousbank/alice-brain Rotary Relay: https://github.com/curiousbank/rotary-relay

The company ultimately chose the lead sales rep with AI experience and Hermes and ended my contract early before the improved system received an evaluation. That experience clarified WIZWIKI’s purpose: not a general-purpose bot, but an owned, teachable marketing platform built around approved knowledge, structured evaluation, distributed private infrastructure, auditability, and human authority. Our goal is one system a company can run, improve, scale, and trust.

Built With

Share this project:

Updates