Auto Alpha AdvisoryThree instruments, each shown with its real status — the live audit engine that reads this page, the AI-visibility tracker, and Komply, in build — and beneath them, the six-system build ledger. No slideware, no softened claims — what’s live is live, what’s sample is labelled, what’s in build says so.
Client case studies ship only once the client has reviewed the numbers and approved the write-up. Until then the evidence here is self-proof: what we’ve built, and our own live engine read — weak rows and all.
Software we run every day, not a portfolio of logos. Each one looks different because each one does something different — and each carries its honest status. As products — what each one is, who it’s for, and where to start — they live on the products page.
Nine domains scored 0–100 with a written narrative — the engine behind your free audit. Run it on your site →
Your share of answer across ChatGPT, Claude, Gemini and Perplexity, swept monthly. Built and running on our own tenant — the graph is illustrative; your real read replaces it. See the AIV Index ↗
in private build · pre-revenue
A South African compliance OS — POPIA, B-BBEE, SARS and FSCA monitoring in one dashboard. In private build, pre-revenue — shown honestly, not yet a live product.
Before AAA sold this work, we built it — for ourselves, and inside the businesses Matt has run and served. Every system below is real, in use, and carries its honest status — the deeper ones link to their build logs.
self-proof · autoalphaadvisory.co.za · read 2026-09-27
AAA’s own engine read returned an overall score of 88 / 100 (Good) on 2026-09-27, across nine scored domains. The read runs on the same audit engine available as the free scan.
The GPTBot SSR test — confirming that the H1, main copy, and JSON-LD are all visible to GPTBot, ClaudeBot, and PerplexityBot in the raw HTML response — passed on the same date.1
Two domains sit lowest. GEO (70) carries the engine’s two high-severity flags on this site: there is no Google Business Profile for AAA, and its entity-name check reads the founder’s name in our structured data as a mismatch with the business. Content (83) reflects two choices we made on purpose: no client case studies until a client has approved the numbers, and a WhatsApp-and-email contact model rather than a published phone line. Everything else reads 88–99.
We publish our own read, weak rows included, from the same engine that will read yours. The read to the right is live, dated, unedited.
audit engine v3 · 9 domains · run 2026-09-27
The law firm read is instructive. Content came back at 92 and Technical at 77 — a site that looks, by those two numbers, like a serious practice with real infrastructure. Then the engine found POPIA at 11 (Critical): the privacy policy was unreachable, in direct breach of POPIA §§ 107–109 (administrative fines up to R10 million and criminal liability up to 10 years).3 GEO — the score that measures how ready a site is for AI search — returned 61: half-ready at best. Forty-nine findings in total. The site looked fine. Nine domains disagreed.
The pattern repeats across all three reads. POPIA is critical in every sector where it was scored. GEO (AI-readiness) is Poor or Fair across the board — not one “Good” in the set. Performance and Email Security are both Poor in every case. Content is the one domain that usually holds up. Everything the machine cares about most is the domain most SA businesses haven’t touched.
Three anonymised real audit-engine runs (2026-06-18). Businesses are labelled by vertical only — no name, no URL.4
See the read before you decide.
The free scan runs the same nine-domain analysis used above on your site, dated and scored, in minutes. The Full Read goes further, into your claims, your terms and who four AI models name when your buyers ask.
The scan is free: no retainer, no commitment. Nine domains, yours.