About Us
LLM Vault is built by John Mackenzie — an AI and cybersecurity specialist who has advised 14 billion-dollar organisations and assessed over US$350 billion in assets across financial institutions, critical infrastructure and government.
Cyber Intelligence. AI Systems. Defensible Decisions.
Who Is John Mackenzie?
John Mackenzie is an AI and cybersecurity specialist, technology developer and independent researcher, based in Australia.
He has spent six-plus years focused on AI and cybersecurity — the years behind the figures above — within a longer career across commercial finance, business and operating technologies. The work has ranged from financial institutions to critical infrastructure and government: organisations where a wrong AI decision is not an inconvenience, it is a reportable event.
An Operator As Well As An Adviser.
John has built and run businesses, not just advised them.
65 Staff In 18 Months
Scaled a digital agency to 65 staff in 18 months.
An 84-Room Hotel
Owned and operated an 84-room hotel — a business run on margins, operations and risk, every single day.
Three Published Books
Authored three published books on cybersecurity and organisational risk.
That operator experience shapes how LLM Vault is built: governance that holds up in a real business, at real pace — not governance that slows the work down.
Credentials And Accreditation.
The depth behind the advice is documented, not implied.
83 Qualifications & Certifications
Across AI, cybersecurity, governance, risk and business strategy.
ISO/IEC 27001 & ISO/IEC 42001
Globally accredited — the international standards for information-security and AI management systems.
Entrepreneurs’ Programme Business Advisor
Australian Federal Government Entrepreneurs’ Programme, 2015–2017.
Two Practices, One Approach.
Through JMacTech and JMac Learning, John helps organisations adopt AI with the governance, oversight and capability to do it safely.
JMacTech
Advisory for organisations adopting AI — the governance, oversight and controls to do it safely, grounded in the assessment work above.
JMac Learning
“Building Capability In An AI-Driven World” — enablement across four domains:
Why Does LLM Vault Exist?
LLM Vault exists because the same gap kept appearing in John’s advisory work: organisations were adopting AI faster than they could govern it.
The industry numbers say the gap is not closing on its own. McKinsey finds 88% of organisations now use AI in at least one business function — yet only around 30% reach adequate governance maturity for agentic-AI controls, just as the era shifts from AI saying the wrong thing to AI doing the wrong thing.
LLM Vault is that approach, productised — the governed layer his advisory work kept proving organisations need. Instead of advice that expires when the engagement ends, the controls run on every AI call, and every decision leaves evidence behind.
LLM Vault is the governed control plane for enterprise AI usage. It sits between an organisation and every large language model it calls — classifying data, gating for compliance and jurisdiction, routing every call to the cheapest model that still meets the requirement, and coaching teams on how well they actually use AI coding tools like Claude Code and Codex. Every decision is logged, every price is sourced, and nothing is trusted just because a client claims it — the platform re-verifies everything itself, on the server, every time. The result is AI usage that is governed, cost-optimised, and auditable by default, not by afterthought.
Talk To John.
Questions about governed AI adoption, the beta, or the approach behind it — email direct, or step into the product.