What We Had to Prove: Inside the Independent Audit Behind Our Microsoft AI Platform Specialisation
- 18 hours ago
- 3 min read
The independent audit, customer validation and technical scrutiny behind the badge.
We just earned Microsoft's AI Platform on Azure Advanced Specialisation. Here's what that required.
Not an exam. Not a self-assessment. Independent auditors reviewed our architectures, interrogated our processes and verified that we've implemented production-grade AI governance across real client engagements, with real consequences if we got it wrong.
The AI landscape is full of announcements. Proofs of concept that never ship. Pilots dressed up as production. Governance frameworks that live in PowerPoints. The AI hype and AI washing are real!
This is a sample of the projects we've built, which have now been independently audited.
Finding the revenue hospitals didn't know they were losing
Incorrect clinical coding costs hospitals money, silently and systematically growing year on year. When costs are assigned to the wrong diagnostic band, revenue from insurers and government payments disappear without anyone realising it.
A major hospital's audit process was manual and slow, and they couldn't identify which cases warranted a human auditor's time due to the volume of data in clinical coding outputs. We built an AI platform that combines OpenAI models, Computer Vision and Azure AI Search to cross-reference coded episodes with clinical notes, surfacing high-value discrepancies and directing expert attention where it matters most.
The result: a proven path to systematic revenue recovery and a working demonstration that AI can make clinical audits smarter without removing human judgment from the equation.
Turning three-week compliance reviews into minutes
A manufacturer was manually checking production specifications and technical documentation against Australian safety and quality standards. Every design revision meant starting again. Errors crept in. Scale was impossible.
We built an automated platform where teams upload technical models and documentation, run checks against the relevant standards, and view violations flagged directly within the design environment. The AI-powered summary feature translates technical findings into plain-language reports for decision-makers, complete with explicit verification prompts, because AI-generated compliance output without appropriate guardrails isn't just unhelpful, it's a liability.
Getting the output framing right was as important as the analysis itself. That's what governed AI looks like.
Putting AI-powered cost intelligence in the hands of customers
A major distributor had a scaling problem. Their business account operations had grown to the point where manual processing was creating friction for customers and internal teams alike. Customers wanted self-service control and real-time visibility over their energy consumption and spending. The internal team wanted to move from transactional support to advisory roles.
We designed and built a comprehensive Azure-native customer portal featuring self-service account management, real-time consumption visibility, spend forecasting, and operational automation, all integrated with billing and back-office systems. The AI piece sits at the heart of the cost intelligence layer: Azure OpenAI Service powers intelligent insights within the platform, helping customers make better decisions about their energy usage and costs rather than simply reporting what's already happened.
For business customers managing consumption-based spend across multiple sites, AI-powered analytics isn't a differentiator; it's table stakes. The governance challenge was ensuring the insights were accurate, explainable, and appropriately scoped. A business customer acting on a misleading consumption forecast has real financial consequences. We built accordingly.
Proving the methodology on ourselves first
We don't just advise clients on AI consulting. We built AIly (AI-ly): our own internal platform that structures discovery work, generates reports, and captures patterns across engagements. Every client engagement we run now benefits from AI that's been tested, iterated, and governed in production.
If you're going to talk about responsible AI implementation, you'd better be able to show it working on your own problems. We can.
Helping Athletes stop leaving money on the table
Athletes consistently undervalue their social media influence, not because it isn't valuable, but because they don't know how to quantify or present it. The Athlete Brand Builder needed to scale that coaching without scaling headcount.
We built Known Athletes: a mobile platform with AI-powered brand profiling, personalised learning pathways, and professional rate card generation that athletes can use immediately in sponsorship conversations. Governing AI outputs that carry financial implications requires accuracy, consistency, and appropriate caveats. Ungoverned AI telling an athlete their audience is worth twice what it is helps nobody.
What the Microsoft AI platform specialisation proved
Across these projects, auditors verified five things: that our guardrails work, that our identity and access controls are tight, that we know where our data comes from, that we red team our systems before clients rely on them, and that we monitor what happens after deployment.
Because AI doesn't stop needing governance once it ships. That's where most implementations quietly unravel. The organisations in these projects couldn't afford to get it wrong: hospitals, vulnerable users, regulated industries, and financial tools. Neither could we.
Read our full thinking on practical AI governance. If you're ready to move from AI experimentation to AI that delivers, let's talk.
