Category: Apps

The Sentient Data Stack: Governing AI-Driven Learning Systems in the Cloud

The Quiet Shift in the Data Stack

Most organisations believe they are adopting AI. However, a more fundamental shift is underway.

AI is no longer confined to isolated models or analytical workloads. It is becoming embedded across the cloud data stack — integrated into pipelines, platforms, and decision layers. As this integration deepens, the role of data systems is shifting from processing information to continuously shaping outcomes.

This is not merely a tooling evolution. It is a behavioural transformation.

Organisations are now building systems that learn from interaction, adapt over time, and influence future decisions based on accumulated feedback. Understanding this shift is critical because it changes not only how data platforms operate, but also how they must be governed.

Defining the Sentient Data Stack

The term “sentient” in this context is used metaphorically. It does not refer to consciousness or awareness. It describes a set of emerging system characteristics:

  • Continuous ingestion of feedback from operational and user interactions
  • Dynamic optimisation of outputs based on observed outcomes
  • Reinforcement loops that strengthen specific behaviours over time

In practical terms, the modern data stack is beginning to behave less like a static repository and more like an adaptive system. Data is no longer simply stored, transformed, and queried. Instead, it is:

  • Feeding models in real time
  • Influencing automated decisions
  • Being reshaped by the outcomes of those decisions

This results in systems that evolve through ongoing interaction with their environment, often within and sometimes beyond initial design parameters — though in practice this evolution occurs inside guardrails and typically requires continuous human oversight in enterprise settings.

From Data Infrastructure to Learning Systems

Cloud data platforms have followed a clear progression:

  • Data pipelines moving information between systems
  • Centralised platforms enabling analytics at scale
  • Integrated ecosystems combining data, models, and automation

The next stage is now emerging. The data platform is becoming a learning system.

In this model:

  • AI is embedded throughout the stack rather than applied externally
  • Decisions are increasingly automated or supported by AI
  • System outputs become inputs that influence future behaviour

Examples of this shift are already visible:

  • Recommendation systems adjusting based on user interaction
  • Forecasting models recalibrating using live operational data
  • Risk systems refining thresholds based on historical outcomes

Individually, these capabilities are well understood. Collectively, they form systems that adapt continuously, often beyond the visibility of traditional governance approaches.

While the building blocks — real-time feature serving, automated model monitoring with drift detection, and feedback ingestion pipelines — are maturing rapidly in major cloud platforms, full autonomous evolution remains gated by risk appetite and regulatory requirements. Most organisations operate in a hybrid mode where learning loops run inside defined guardrails with human oversight and periodic intervention.

Proprietary Data as a Learning Advantage

In this environment, the role of data changes fundamentally. It is no longer simply an asset to be stored and analysed. It becomes the basis of learning capability.

Competitive advantage begins to depend on:

  • The quality of feedback loops rather than volume of data alone
  • The uniqueness of proprietary datasets combined with curation and feedback engineering
  • The speed and discipline with which systems learn and adapt

Two organisations may deploy similar technologies yet achieve different outcomes based on how their data feeds learning cycles and how effectively those cycles are governed. This introduces a new form of differentiation. The critical factor is not who has the best model, but who develops the most effective, well-governed learning system.

Governance in a Learning Environment

Traditional data governance is designed for stability. It is based on:

  • Defined data structures
  • Controlled data flows
  • Policy-based access and compliance

These approaches remain necessary but are no longer sufficient.

In a learning system:

  • Data continuously influences behaviour
  • Models adapt as new inputs are introduced
  • System outputs shape future decisions

Governance must therefore evolve from static control to continuous oversight. This requires:

Embedded Governance

Governance mechanisms must be integrated directly into systems. This includes data lineage tracking, model monitoring, automated enforcement of policies, and outcome tracing.

Real-Time Observability

Leaders require visibility into system behaviour — what the system is doing, how it is evolving, and why decisions are being made. This aligns with the NIST AI Risk Management Framework’s emphasis on ongoing monitoring within the Measure function and adaptive risk treatment in Manage.

Behavioural Oversight

The central question changes from compliance to behaviour. It is no longer sufficient to ask whether a system meets requirements. It is necessary to understand whether it is behaving as intended. A system can remain compliant while producing outcomes that introduce strategic or operational risk.

Practical mechanisms include drift detection (data, concept, and performance), continuous evaluation harnesses, and automated policy-as-code enforcement extended to AI decision points. These are particularly critical in public sector and regulated environments, where decisions must remain explainable, contestable, and aligned with administrative law principles.

The Risk of Self-Reinforcement

The most significant risk in a learning system is not failure. It is uncontrolled reinforcement.

Feedback loops strengthen patterns over time. When those patterns include bias, misaligned objectives, or incorrect assumptions, they are amplified. This can result in:

  • Bias in decision-making
  • Operational drift away from intended outcomes
  • Strategic misalignment driven by incorrect optimisation signals

These risks are difficult to detect. They do not present as system failures. They emerge as gradual shifts in behaviour that become embedded over time.

A prominent example in generative contexts is model collapse, where recursive training on synthetic or AI-generated outputs leads to progressive degradation in performance, diversity, and factual reliability. The same underlying dynamic — outputs becoming inputs without sufficient controls — can appear in any closed-loop decision system.

Managing these dynamics requires a shift in mindset. Leaders are no longer focused solely on preventing failure. They are responsible for controlling how systems evolve.

Trust as an Architectural Layer

Trust cannot be treated as a compliance outcome. It must be engineered into the system. This requires:

  • Data lineage, providing visibility into the origin and movement of data
  • Decision traceability, enabling explanation of outcomes
  • Model transparency, offering insight into system behaviour and change

Trust becomes a system capability. It supports confident decision-making, regulatory alignment, and organisational adoption of AI-driven systems. Without it, systems may function, but they will not be relied upon.

Executive Implications

For executive leaders, this shift introduces new responsibilities. The focus moves from managing systems to governing behaviour.

Key questions include:

  • What is the system learning over time?
  • Which signals are reinforcing its behaviour?
  • Where is accountability defined?
  • How transparent are its decision-making processes?
  • What is our organisation’s capacity for continuous behavioural assurance?

These are not technical considerations. They are strategic governance issues that determine whether an organisation is building a controlled learning system or allowing one to evolve without sufficient oversight.

The Governance Mandate

The sentient data stack is not a future concept. Its components are already present in modern cloud environments, shaped by embedded AI capabilities, real-time data integration, and continuous feedback mechanisms.

The shift is underway. The question is how effectively it will be governed.

Organisations that succeed will not be those that deploy the most AI. They will be those that:

  • Understand how their systems learn
  • Establish control over feedback loops
  • Build trust into system architecture from the outset

In doing so, they move beyond managing data and begin governing behaviour at scale.

Resources 

The perspectives outlined in this article are informed by observable industry developments and established frameworks relating to cloud platforms, AI integration, and governance. 

Key reference sources include: 

NIST Artificial Intelligence Risk Management Framework (AI RMF)

NIST Data Governance and Risk Management Publications

Australian Cyber Security Centre (ACSC)

CSIRO – Australia’s National Science Agency

Cloud Platform Architecture and Governance Guidance 

Industry Analysis and Strategic Research 

From Reactive AI to Agentic Systems: The Rise of Goal-Driven Intelligence in the Cloud 

AI Was Never the End State 

For years, organisations have invested heavily in artificial intelligence, building capabilities around chatbots, predictive models, and recommendation engines that have steadily delivered value across different parts of the business. 

These systems have been effective, but only within clearly defined boundaries. They respond to inputs, analyse data, and automate tasks that have already been mapped out in advance. 

What they do not do is think ahead, plan independently, or act with intent beyond the instructions they are given. 

That distinction matters more now than it ever has, because the role of AI is starting to shift. 

The Limits of Reactive AI 

Most AI systems deployed today are still fundamentally reactive in nature. They rely on a simple cycle: wait for input, process the data, and return an output. 

This model works well in controlled environments and continues to support use cases such as customer support automation, forecasting, analytics, and content generation. 

However, as operating environments become more dynamic and interconnected, the limitations of this approach become increasingly visible. 

Reactive systems struggle to orchestrate multi-step processes, adapt strategies in real time, coordinate across multiple systems, or act without explicit prompts. These gaps are not due to a lack of intelligence, but rather a lack of agency. 

And that is where the real constraint lies. 

Enter Goal-Driven AI Agents 

Agentic AI introduces a fundamentally different operating model. Instead of waiting for instructions, systems are designed to operate against defined objectives and take the necessary steps to achieve them. 

This means they can break down goals into smaller tasks, determine which tools or data sources are required, execute actions across systems, and continuously evaluate outcomes to refine their behaviour. 

In practical terms, this could involve a system monitoring a supply chain and adjusting inventory levels before disruptions occur, or managing a marketing campaign that optimises itself across multiple channels without constant human input. 

It might also include identifying inefficiencies within internal systems and triggering optimisation workflows, or coordinating different tools to complete complex tasks from start to finish. 

The shift may appear subtle on the surface, but it represents a meaningful change in how systems operate. 

We are moving from a model that responds to instructions toward one that actively pursues outcomes. 

Why the Cloud Is Critical 

Agentic AI does not operate in isolation, and its effectiveness is closely tied to the capabilities of the cloud environments in which it runs. 

These systems depend on continuous access to data, scalable compute resources for reasoning and execution, seamless integration across APIs and enterprise platforms, and real-time feedback loops that allow them to adjust behaviour as conditions change. 

Without this underlying infrastructure, it becomes difficult to orchestrate workflows across systems, scale decision-making processes, or maintain the persistent context required for autonomous operation. 

This is where previous cloud investments begin to deliver compounding value. Cloud is no longer just the environment in which AI is hosted; it is the foundation that enables autonomous systems to function at scale. 

Frameworks Enabling Agentic AI 

The rise of agentic AI is being accelerated by a new generation of frameworks designed to support orchestration, memory, and multi-step execution. 

LangChain, for example, allows developers to connect language models with tools, memory, and workflows, enabling systems to maintain context across interactions and execute more structured processes. 

CrewAI extends this further by introducing multi-agent collaboration, where different agents take on specific roles and work together toward a shared objective, creating a system that begins to resemble a coordinated digital workforce. 

Emerging frameworks such as OpenClaw point toward a more flexible and open approach to agent orchestration, where organisations can design and customise how agents behave rather than relying solely on pre-defined capabilities. 

This reflects a broader shift in expectation. Organisations are no longer just looking for powerful models; they are looking for systems they can shape and control. 

What This Means for Business Operations 

The introduction of agentic AI is not simply an incremental improvement in efficiency. It represents a structural shift in how work is executed within organisations. 

The traditional model of humans interacting with tools to produce outputs is gradually being replaced by a model where humans define goals, and systems take on the responsibility of executing toward those outcomes. 

This has direct implications across multiple areas. 

Operationally, routine coordination tasks can become autonomous, allowing teams to focus more on direction and strategy. 

In decision-making, AI moves beyond providing insights and begins to act on them within defined parameters. 

From a productivity standpoint, the nature of work shifts from task execution to system oversight, while scalability improves as organisations can expand operations without a corresponding increase in headcount. 

At the same time, this shift introduces a new layer of complexity. 

The Governance Challenge 

As systems gain more autonomy, the importance of governance increases significantly. 

Organisations must define what decisions AI agents are allowed to make independently, establish clear boundaries, and ensure that actions can be audited and traced when needed. 

Questions around accountability also become more prominent, particularly in situations where systems are making decisions that have real operational or financial impact. 

Agentic systems have the potential to amplify both capability and risk. Without the right governance structures in place, increased autonomy can quickly translate into increased exposure. 

This is where architecture, policy, and cloud infrastructure need to work together as a cohesive system. 

Strategic Checkpoint 

For organisations exploring this space, it is worth taking a step back and assessing readiness from a broader perspective. 

Are your systems designed to execute, or are they still primarily focused on analysis? 

Can your infrastructure support continuous, autonomous workflows? 

Do you have governance models in place to manage AI-driven decision-making? 

And are you actively experimenting with agentic systems, or still relying solely on prompt-based interactions? 

These are not theoretical questions. They are practical considerations that will shape how effectively organisations can adapt to what is already unfolding. 

Final Thought 

AI is no longer just a tool that supports isolated tasks. It is becoming a system of action that influences how work is carried out across the organisation. 

The transition from reactive models to goal-driven agents marks a significant shift in how technology contributes to business outcomes. 

However, the real advantage will not come from adopting these systems alone. It will come from designing the environments in which they operate, ensuring that they are secure, governed, and aligned with organisational objectives. 

Because ultimately, the value of AI will not be measured by what it can say. 

It will be measured by what it can do. 

Resources 

All views are my own personal opinions.

Stay informed on the evolving intersection of cloud, AI, and digital transformation.

Subscribe to the newsletter for monthly insights, strategic analysis, and emerging trends shaping enterprise and government technology.

Public Sector Cloud Strategy and Transformation in Australia

Are Australian governments unlocking the full cloud promise; or just testing the waters? 

Australia’s public sector has embraced cloud as a strategic enabler, not just an IT upgrade. But with rising expectations for secure, citizen-centric digital services, agencies now face a tougher challenge: moving from cloud adoption to cloud maturity. 

The ambition is high. The complexity is real. And the decisions government leaders make now will shape the next decade of public sector capability. 

Embedding Hybrid & Public Cloud into the National Digital Government Vision 

Australia’s Data and Digital Government Strategy (2023) and the forthcoming Whole-of-Government Cloud Computing Policy (effective July 2026) mark a pivotal shift: cloud-first is no longer aspirational. It’s the default operating model for modern government. 

This isn’t just a technology upgrade. It’s a structural change in how the government delivers services, manages risk, and builds resilience. 

These frameworks push agencies to: 

  • Use public cloud for new digital services 
  • Actively retire legacy and high-risk systems 
  • Prioritise reusable, interoperable platforms 
  • Modernize procurement and governance 
  • Strengthen whole-of-government consistency 

In short, we’re moving from siloed ICT to shared national digital infrastructure, a foundation that supports collaboration, agility, and citizen trust. 

Hybrid cloud plays a critical role in this vision. Public cloud accelerates innovation and scalability, but hybrid architectures allow agencies to keep sovereignty over sensitive workloads while still gaining flexibility and cost efficiency. It’s not about choosing one or the other. It’s about designing a model that balances speed with control. 

This shift raises a leadership question: 
Does your Digital Investment Plan treat cloud as infrastructure or as a strategic capability that shapes service delivery? 

Secure Cloud Strategy: Building Resilience, Assurance & Agility 

The Secure Cloud Strategy, supported by ASD’s Blueprint for Secure Cloud, is designed to move the public sector away from “lift and shift” thinking toward secure design. 

This isn’t just a checklist for compliance. It’s a mindset shift. From treating cloud as a convenient hosting option to recognizing it as part of Australia’s critical national infrastructure. 

The strategy provides practical tools: 

  • Architecture patterns for secure deployment 
  • Risk assessment templates 
  • Guidance for configuration and hardening 
  • Clear shared responsibility boundaries 
  • Controls aligned to the PSPF and Essential Eight 

But the real goal is bigger … modernizing how government builds, tests, and operates services in an environment where threats are constant, and citizen expectations are uncompromising. 

A secure cloud strategy must help agencies: 

  • Detect threats faster 
  • Mitigate failures gracefully 
  • Respond to crises without downtime 
  • Maintain high trust in public digital services 

Cloud is no longer just a place to store applications. It’s the backbone of the digital government. Every outage, every breach, every delay impacts public confidence. Treating cloud as critical infrastructure means designing resilience, agility, and assurance from day one. 


Is your cloud security posture reactive or embedded as a strategic capability that protects trust and continuity? 

Navigating Sovereignty, Security & Complexity 

Australia’s cloud environment isn’t just shaped by technology choices, it’s defined by regulatory guardrails, security expectations, and sovereignty obligations that few other markets face. For public sector leaders, these aren’t optional considerations. They’re foundational to trust and compliance. 

Data Sovereignty: More Than a Location Requirement 

Under the PSPF, Privacy Act, and sector-specific rules like the My Health Records Act and APRA CPS 234, agencies must ensure sensitive data: 

  • Stays within Australian jurisdiction 
  • Is processed by vetted and accredited providers 
  • Aligns with sovereign risk and resilience standards 

This is why sovereign cloud regions, such as those in Canberra and Sydney, matter. They’re not just technical zones. They’re protected environments for workloads with national sensitivity, ensuring that critical data remains under Australian control. 

Across Australia’s cloud policy, protective security framework, and secure cloud guidance, the message is consistent: sovereignty is a foundation for resilience. Control over data, jurisdiction, and access is not about geography alone; it is about reducing national risk, strengthening security posture, and ensuring continuity in the face of disruption. 
Read more here. 

Security Expectations: Continuous, Not Occasional 

Australia’s Essential Eight maturity model, ISM controls, and PSPF frameworks demand more than periodic audits. They require ongoing posture management, because in a cloud world, risk is dynamic. 

That means: 

  • Continuous monitoring 
  • Policy automation 
  • Zero Trust architecture 
  • Governance at scale 

Security isn’t a bolt-on. It’s a living capability that evolves as threats evolve. 

Operational Complexity: The Hidden Challenge 

Cloud promises simplicity, but reality often looks different. Agencies face: 

  • Multi-cloud governance friction 
  • Cost unpredictability 
  • Talent shortages for cloud-native skills 
  • Risk of over-dependence on a single vendor 

Recent ANAO audits show that failures rarely stem from cloud itself. They come from governance, maturity, and lagging adoption. Technology moves fast. Policy and capability must be kept at a pace. 

 
Is your agency treating sovereignty, security, and complexity as compliance hurdles or as strategic levers for trust and resilience? 

Sharpening Cloud ROI & Agility: CIO Best Practices 

Cloud maturity isn’t measured by how many systems an agency migrates. It’s measured by how effectively cloud supports outcomes, resilience, cost efficiency, service improvement, and risk reduction. The question isn’t “How much cloud do we have?” but “How much value does it deliver?” 

High-performing CIOs in the public sector share a common approach. They treat cloud as a strategic capability, not just infrastructure. Here’s what sets them apart: 

1. Strategic Governance Built-In 

Cloud strategy must be embedded early in Digital Investment Plans, not bolted on later. Governance isn’t paperwork. It’s the guardrails that keep transformation on track. 

What does this look like? 

  • Portability clauses to avoid lock-in 
  • Vendor-neutral patterns for flexibility 
  • Reusable reference architectures 
  • Clear multi-cloud guardrails 

This ensures consistency across agencies and reduces reinvention. It’s about building a system that scales without chaos. 

2. Cost Transparency & Control (FinOps for Government) 

Cloud can be a silent cost escalator if left unchecked. That’s why the government is adopting FinOps disciplines, blending finance and operations to make spending visible and accountable. 

Key practices include: 

  • Real-time monitoring across providers 
  • Workload right-sizing 
  • Clear unit costing 
  • Independent audits of cloud use 

The goal? Every dollar spent on cloud should map to measurable public value. 
Read the FinOps Public Sector Whitepaper. 

3. Agile, Risk-Aware Security Models 

ASD’s blueprint stresses continuous, adaptive security. CIOs must: 

  • Align provider responsibilities 
  • Automate compliance checks 
  • Standardize configurations across environments 

Security isn’t a static policy. It’s a living system that evolves as threats evolve. 

4. Effective Hybrid Architecture 

Sensitive workloads often remain in sovereign regions or protected private environments, while scalable digital services leverage public cloud elasticity. The challenge? Integration. 

Legacy systems and modern cloud-native platforms must interoperate seamlessly, securely, reliably, and under consistent governance. This is where architecture discipline meets operational reality. 

5. Culture, Skills & Centers of Excellence 

Technology transformation fails without workforce capability. Agencies benefit from creating: 

  • Cloud Centers of Excellence (CCoE) 
  • Cloud-native training pathways 
  • Shared learning across government 
  • Communities of practice 

This builds consistent standards and accelerates adoption. Cloud isn’t just a tech shift. It’s a cultural one. 

6. Measuring Business Outcomes 

CIOs are moving beyond technical KPIs to outcome-based metrics: 

  • Reduced operating risk 
  • Improved citizen experience 
  • Strengthened service resilience 
  • Shorter delivery cycles 
  • Lower cost-to-serve 

Cloud success is strategic, not technical. It’s about impact, not infrastructure. 

 
Is your cloud program measured by migration milestones, or by the outcomes that matter most to citizens and government resilience? 

Looking Ahead: The Cloud-Enabled Public Sector 

Australia is building a public sector cloud ecosystem that balances innovation with sovereignty, resilience, and trust. 

The next step is consolidation. Not just running hybrid environments, but aligning them into a cohesive, cloud-native platform for the entire public sector. 

The real test: 
Can agencies deliver unified, citizen-first public services while managing risk, cost, and national control? 

 
What would success look like for your agency’s cloud-first journey in 2026 and beyond? 

Resources 

• Data and Digital Government Strategy (DTA) 

https://www.dta.gov.au/our-initiatives/data-and-digital-government-strategy

• Whole-of-Government Cloud Computing Policy 

https://www.digital.gov.au/cloud-policy

• Secure Cloud Strategy 

https://architecture.digital.gov.au/strategy/secure-cloud-strategy

• ASD Blueprint for Secure Cloud 

https://blueprint.asd.gov.au/

• Cyber.gov.au – Cloud Computing Guidance 

https://www.cyber.gov.au/business-government/protecting-devices-systems/cloud-computing

• Protective Security Policy Framework (PSPF) 

https://www.protectivesecurity.gov.au/

• Privacy Act 1988 

https://www.oaic.gov.au/privacy/privacy-legislation/the-privacy-act

• Digital Investment Management Framework (DIMF) 

https://www.dta.gov.au/our-initiatives/digital-investment-management

• Australian Government Architecture – Cloud and Hosting 

https://architecture.digital.gov.au/domains/cloud-and-hosting

• ASD Essential Eight 

https://www.cyber.gov.au/resources-business-and-government/essential-cyber-security/essential-eight

• ANAO Reports and Audit Insights 

https://www.anao.gov.au/

• FinOps Framework (FinOps Foundation) 

https://www.finops.org/framework/

Cloud in Healthcare: How Australia is Using AI to Transform Digital Health 

What if your next medical breakthrough isn’t a new drug or device— 
but the cloud infrastructure running quietly behind the scenes? 

Australia’s healthcare system is undergoing a quiet revolution. And at the heart of it isn’t just AI, or machine learning, or cutting-edge telehealth tools—it’s the rapid evolution and reach of cloud computing. 

From telemedicine in remote towns to real-time hospital analytics in the CBD, cloud infrastructure is no longer an IT decision. It’s a care decision. And it’s accelerating faster than most organisations are ready for. 

The Rise of Cloud in Australian Healthcare 

Cloud computing in Australian healthcare has gone from experiment to essential. 

In 2022–23, 20% of all GP services were delivered via telehealth—phone and video are now a standard part of care delivery, particularly in rural and aged care settings. 

Electronic Health Records (EHRs) are evolving from static repositories to dynamic, AI-ready platforms. 

Predictive analytics is helping hospitals forecast admissions, manage resources, and reduce waiting lists. 

But with every new capability comes a challenge: integration, security, governance, and compliance. 

Cloud has shifted from a back-end technology to a strategic engine for growth and innovation. It’s becoming the backbone of modern health delivery—and the risk and compliance surface has expanded accordingly. 

AI in Action: Smarter, Faster, Fairer Care 

Australia is at the forefront of AI and ML innovations in healthcare. 

  • AI triage bots are helping assess symptoms and direct patients to appropriate care pathways. 
  • Machine learning models are predicting patient deterioration in emergency rooms. 
  • Natural language processing is accelerating clinical documentation, giving practitioners more time with patients. 
  • Computer vision is assisting radiologists in detecting anomalies more quickly and accurately. 

These use cases are not hypothetical. They are operational today—and they rely on scalable, secure cloud environments. 

However, these technologies are only as strong as the infrastructure they run on. And in healthcare, that infrastructure must meet an exceptionally high bar. 

The Privacy and Compliance Tightrope 

Healthcare cloud adoption in Australia must navigate a complex environment of privacy laws, ethical obligations, and system-wide compliance expectations. 

Technology teams supporting healthcare are not simply managing digital records—they are stewards of public trust. 

The Privacy Act 1988  and the My Health Records Act 2012  impose clear responsibilities around data sovereignty, consent, and transparency. 

The Australian Digital Health Agency maintains national standards for interoperability, access controls, and cybersecurity. 

Accreditation frameworks such as ISO/IEC 27001  and IRAP (Information Security Registered Assessors Program) are becoming mandatory in procurement processes. 

Choosing the wrong cloud partner is not just a technical oversight. It becomes a compliance issue, a reputational risk, and an ethical liability. 

Choosing the Right Cloud Partner for Healthcare in Australia 

For healthcare leaders, selecting a cloud partner in healthcare is no longer a purely operational decision—it is a strategic one. 

At a minimum, ensure your cloud solution offers: 

  • Data residency within Australia 
  • IRAP-assessed infrastructure 
  • Proven interoperability with national digital health systems 
  • Capacity to support AI and machine learning workloads 
  • Transparent security protocols, SLAs, and audit trails 

Above all, choose a partner who understands that in this sector, the goal is not disruption. The goal is safe, sustainable, patient-focused innovation. 

Final Thought 

If you’re leading technology in a healthcare organisation, the question is no longer whether cloud and AI should be adopted. 

The real question is: are we building the kind of infrastructure that can support the next decade of health innovation? 

Because in the end, this is not just about platforms and data. It is about empowering clinicians. It is about faster, more informed decisions. And ultimately, it is about improving lives—quietly, securely, and intelligently in the background. 

Let’s build that future—thoughtfully, together. 

Resources 

1. MBS Telehealth Post-Implementation Review Final Report 
https://www.health.gov.au/sites/default/files/2024-06/mbs-review-advisory-committee-telehealth-post-implementation-review-final-report.pdf 

2. Patient Experiences in Australia 
https://www.abs.gov.au/statistics/health/health-services/patient-experiences/latest-release 

3. Australia Telehealth Market Report 2025–2034 
https://www.expertmarketresearch.com.au/reports/australia-telehealth-market 

4. Privacy Act 1988 
https://www.oaic.gov.au/privacy/privacy-legislation/privacy-act-1988 

5. My Health Records Act 2012 
https://www.legislation.gov.au/Details/C2012A00184 

6. IRAP – Information Security Registered Assessors Program 
https://www.cyber.gov.au/acsc/view-all-content/programs/irap 

7. ISO/IEC 27001 – Information Security Management 
https://www.iso.org/isoiec-27001-information-security.html 

8. FHIR (Fast Healthcare Interoperability Resources) 
https://www.hl7.org/fhir/ 

9. Real-Time AI for Patient Deterioration Prediction

Source: National Library of Medicine (PubMed)

https://pubmed.ncbi.nlm.nih.gov/37150397/

10. AI Chatbots in Australian Healthcare

Source: University of Melbourne, Pursuit
https://pursuit.unimelb.edu.au/articles/the-promise-and-peril-of-ai-chatbots-in-healthcare

11. Computer Vision in Radiology (SA Medical Imaging)

Source: Adelaide Now (News Corp Australia)
https://www.adelaidenow.com.au/news/south-australia/artificial-intelligence-advising-on-xray-diagnoses-in-sa-medical-imaging/news-story/ae20cc4c30320354069d586ca1d23846

Publishing a .NET Core App

In my last blog, I mentioned a basic Web App that I had put together using Visual Studio 2015 tooling for .NET Core. Over the last couple of days, I’ve been looking at publishing the App, and the steps involved.

.NET Core is a new beast with a lot of potential…. the aim is it will run anywhere, on anything. So to keep my costs down, I’m going to trial it on my Amazon Linux VM. Note to self, SQL Server for Linux is about 1 year down the track – unfortunate, as the MVC scaffold uses SQL Server LocalDB – so I’ll have to figure out what database I can use.

But, first things first… how to publish my App to Linux?

First step – publish the App using dotnet publish. I found that Bower was not referenced in my Path environment variable. Bower was installed with Visual Studio 2015 Professional – in a sub-folder of my Visual Studio 2015 folder. I’m not sure if it was installed because I had installed tooling for .NET Core, or if it comes by default. Anyway, once that was sorted, dotnet publish worked fine and it created a portable for me to use.

I copied all of the files in the folder that dotnet publish created over to my Amazon Linux server. (I used WinSCP for this). Then I found I needed to install .NET Core on Amazon Linux. Installation was easy but when I tried to run the dotnet CLI, I received an error. Running dotnet –info from a bash shell I saw,

dotnet: /usr/lib64/libstdc++.so.6: version `GLIBCXX_3.4.18′ not found (required by dotnet)

After several hours of searching, I found this was due to the libstdc++ library version on my Amazon Linux distribution, which was libstdc++47. I had ensured my VM was up to date, so it seems that the libstc++ version was lagging (for whatever reason). After running the below, I was able to successfully run the dotnet CLI, with a valid response from dotnet –info.

sudo yum install libstdc++48

So, I had installed .NET core and fixed up the reference library it needed. After that, I needed a way to access my Web App. Kestrel is a web server built into .NET Core and can listen on any port you tell it to. However, it is better to use a web server as a proxy/reverse-proxy to relay request/response to Kestrel.

I already had a web server on my Amazon Linux VM. So I configured it with proxy and reverse-proxy mappings (to the Kestrel server in my .NET Core App). The calls to my web server forward on to the Kestrel server in the App, and vice-versa. I ran the App with dotnet run and checked access via the web server. All good 🙂

Almost there.

Finally, I installed supervisor to manage start/stop/restart of my App. And set up a script to ensure supervisor is re-started whenever my Amazon Linux machine restarts.

Note, this would have taken days if not for the early groundwork of several people who blog their efforts.

Now I have a Web App but no database. So next step is to get the database up and running.

 

Scrapy Spiders, Python Processing & Web APIs

Over the past couple of weeks, I’ve spent some time drafting a Web App for Touch Footy results. The App is built on .NET Core, and this gave me a great opportunity to review the new Visual Studio .NET Core tooling. But once I had my bare bones App, I needed some data to play with. Enter Scrapy and Python…

What I wanted was a data set that I could use in the Touch Footy App. I had a good data source and I figured my best bet was a web scraper. Scrapy made it easy for me to scrape together my test data set. It’s built using Python (hence you need some understanding of Python to use it). Python is an interpreted language. It’s great for list processing and it’s easy to read/write.

Scrapy is an open source framework for writing web crawlers, or spiders. It gives you control over how and when you execute the spiders you’ve written. And a great shell as a part of the framework to test/debug commands. After looking at other web scraping options, I decided on Scrapy as a neat way to get my data.

After a few hours coding, I had a crawler that collected the data I wanted, i.e. groups, teams, fixtures and results for my Web App. I wanted to store the data in JSON – for easy processing – and Scrapy made that easy too. It was simple then to write some Python code to process the JSON for groups, teams, fixtures and results.

All good so far, and fun to boot. My next step – how to get the data to the Touch Footy Web App? Well, Visual Studio 2015 tooling for .NET Core makes it easy to add a Web API to an MVC Web Application…. several hours later I had a working spider populating data into my App.