A real snapshot, not a polished pitch — what's active, what's on hold, what's still just a concept, and every live tool with a real link. Updated 2026-09-24.
Stated long-term goal (Michael, 2026-09-02): $1 billion within 10 years. Everything below is a real, parallel-tracked step toward that — Michael works multiple fires at once by design, not as a focus problem.
AI automation retainer for local trades businesses, built on GoHighLevel
A$999/month flat retainer covering AI voice receptionist, review automation, missed-call text-back, booking calendar, unified inbox, social planner, and quote automation — delivered white-label through GoHighLevel, never shown to the client. Real domains purchased (exvoria.com.au, exvoria.au). First real prospect is Michael's cousin (wedding venue/flower farm/podcast) — a tailored assessment report has been built for that pitch, not yet confirmed sent.
AI Governance & Readiness for NFP/Charity Boards
Advisory/coaching model, not GHL automation — a genuinely different delivery shape from Vertical 1, same brand. Grounded entirely in the real AICD/HTI Director's Guide to AI Governance, ACNC's own sector guidance, and a 2026 Federal Court judgment (ASIC v Bekier) where a judge addressed "Shadow AI" directly. Positioned as a real study you and Michael are running, not a sales pitch — that's the frame this whole vertical is built on.
Michael Szafraniec, publicly — the leverage lever, not just Exvoria's founder
Built off Daniel Priestley's personal-brand framework (7-11-4, Name-Same-Fame-Aim-Game, free value beats scarcity, discussion groups). Reasoning: one-to-one consulting is bell-curve-bound by hours; a personal brand is the one asset here that compounds without more hours. The AI governance book above is the flagship free-value piece.
Gumroad + Etsy digital product lines, sitting alongside the main service businesses
15 digital products are live on Gumroad right now (Exvoria side), including an NDIS & Aged Care Compliance Kit (A$39, built around the real 2026 reforms — 7-year record retention, care-minute accountability) linked from a dedicated Aged Care/NDIS positioning page. Zero sales on any of them yet — the honest next action is posting the already-queued announcement content, not building more products. Separately, hearthomelife has 10 Etsy digital products planned (cleaning schedule bundle, home inventory tracker, kitchen pantry set, new home checklist, seasonal maintenance calendar) — content and Canva prompts exist, nothing's actually been built in Canva or listed on Etsy yet. A hearthomelife print-on-demand shop and a separate, unrelated novelty t-shirt POD brand are also planned, part of the same side-hustle sprint as the newsletter.
Hex's own internal multi-agent setup, modeled on real enterprise agentic-workforce patterns
Seven role-scoped agents exist as real, invocable Claude Code commands (not personas): a Chief of Staff that orchestrates the other six, plus Exec Assistant, CMO, CFO, COO, Risk Manager, and a Monitor that cross-checks the other five for drift or unverified claims. Each reads its own charter and accumulated memory, does one bounded task against real project state, and reports back. Honest caveat, not glossed over: this has not been proven reliable at scale — a real smoke test hit an actual Anthropic API credit-balance issue and a pm2-hanging bug in a headless session, both root-caused and fixed, but the roster is still "built and wired," not "battle-tested." A `local-lead-gen-scorer` skill also exists for Website Landlord's niche scoring. No project-level MCP config exists — live integrations (Notion, etc.) are session-level, not a fixed file.
Home goods affiliate store (Shopify) — beeswax candles, cushions, lighting, home accessories
Amazon AU + US affiliate links, blog + social content pipeline. Separately, a side-hustle sprint is in progress: a real-career newsletter and a print-on-demand line running through the existing HHL Shopify store, plus a second, separate novelty-t-shirt POD brand — both aimed at generating revenue faster than the affiliate model alone.
Exact-match-domain local-service lead-gen sites, ranked organically, leads rented (not sold) to a real local business
Build once, rank for ~2 months, then rent the leads it generates — the site itself never changes hands. Pipeline (niche/geo scorer, content research, multi-page site generator, weekly rank-checking) was built and tested end-to-end. Regular ranking reports are actively being generated day to day, though I haven't done a fresh deep-dive on this specific venture's current live-site count tonight — worth a direct check with Michael if you want the precise current number.
Helping NFPs deploy idle cash reserves into property that houses Community Housing Providers / NDIS accommodation
Michael would operate as a buyers agent (Cert 4 qualified, needs a licensed principal or exemption pathway) — commission on acquisitions, advisory retainers, cooperative structuring for smaller NFPs pooling capital. Nothing built yet; next real step would be a landing page and a shortlist of 5 target NFPs.
Civic intelligence platform — evidence-based decision support for governments, NGOs, and communities
Deliberately not being built yet. The standing rule: "built properly or not at all — credibility is the product." Waiting on Exvoria's revenue to cover its own operating costs before Solera becomes fundable. A real number was set 2026-09-21 — Exvoria needs $150k/month — and the honest math says the current combined ceiling across both verticals, as structured, is closer to $20-30k/month. That gap is a known, open, unsolved problem, not something quietly dropped.