Utilising AI for Business Applications

This module shows you the difference between genuinely changing how work happens and simply bolting a tool onto the side, and it gives you an honest way to think about return on investment for a small or medium business. You will learn where to start, and you will learn the failure modes that quietly waste everyone's time so that you can steer around them.

• 01

  The 92/19 gap: why most AU SMEs are not seeing returns

  You will understand why almost every Australian small and medium business is using AI while only a small fraction have actually changed a single workflow, and you will see the concrete patterns that close that gap inside a real business.

• 02

  Value mapping: what is worth automating, what is not

  You will learn to map your work along two dimensions, how often a task happens and how much human judgement it needs, so that you can tell at a glance what is worth automating first and what is genuinely better left alone.

• 03

  ROI for AI, properly

  You will see why so many Australian organisations cannot measure the return on their AI spend, and you will come away with a way of framing return on investment that actually holds up for a small or medium business.

• 04

  Starting points that work (and ones that do not)

  You will learn to tell a promising first project from a risky one, and you will be able to pick a pilot using a simple filter that favours work which happens often and where a mistake would do little damage.

• 05

  What NOT to do

  You will learn the five most common ways Australian small and medium business rollouts come undone, and you will be able to spot each of them in your own plans before they cost you anything.

Here we look properly at Claude, ChatGPT, Gemini and Copilot for real business use, covering where your data goes, how to run inference within an Australian region on Bedrock in Sydney or Azure in Australia East, and how to budget sensibly. You will also get a vendor procurement checklist and a clear sense of when sovereign options genuinely matter.

• 01

  Claude vs ChatGPT vs Gemini vs Copilot, for business

  You will compare the four major options at the depth a business actually needs, weighing security, compliance, integration, and what each one really costs at the scale of a small or medium business.

• 02

  Data residency: where your data goes and why it matters

  You will understand your cross border disclosure liability under APP 8, what it means to run inference within an Australian region, and how to recognise the situations where keeping data onshore is a genuine hard requirement rather than a nice to have.

• 03

  The AU vendor procurement checklist

  You will walk away with a downloadable checklist of roughly twelve to fifteen questions that every Australian small or medium business should put to an AI vendor before signing anything.

• 04

  Budgeting AI for an SME

  You will learn the difference between paying per seat and paying per token, and you will see a realistic twelve month spending pattern for a business of five to twenty five people so that the costs hold no surprises.

• 05

  Sovereign and Australian-made options

  You will get a clear view of the Australian built and Australian sovereign options available to you, and you will learn to tell when being Australian made genuinely matters and when it is simply marketing.

This module treats your system prompts as company policy and walks you through knowledge bases, personas, and a sensible way to version and audit what you have built. You will see the eight Australian AI Ethics Principles applied to the kinds of customisation choices you will actually make.

• 01

  System prompts as company policy

  You will learn to write a system prompt that quietly encodes how your team works, setting the tone, the role, the way things escalate, and the formatting standards everyone should follow.

• 02

  Knowledge bases and document libraries

  You will learn how to use Claude Projects, ChatGPT Custom GPTs, and Copilot Studio to build a knowledge base, and just as importantly you will learn what belongs in it and what is better left out.

• 03

  Personas and roles

  You will learn when a single persona is not enough, see the personas that come up again and again in small and medium businesses, and be able to write one that the AI will actually follow.

• 04

  Versioning and audit (without enterprise tooling)

  You will set up a simple way to track what is running and why, giving you a level of audit logging that is realistic for a small or medium business without any enterprise tooling.

• 05

  The eight AI Ethics Principles in practice

  You will work through each of the eight Australian AI Ethics Principles at a practical depth, with worked examples that genuinely change the decisions you make.

You will learn the real distinctions between a workflow, an autonomous agent, and a multi agent setup, and you will work through the six configuration layers that sit at the heart of the Resonance approach. By the end you will have three agent specifications concrete enough to hand straight to a developer.

• 01

  Workflow vs autonomous agent vs multi-agent

  You will learn the real distinctions between a workflow, an autonomous agent, and a multi agent setup, weighing the trade offs in cost, risk, and capability, and you will finally have a clear answer to whether something is truly an agent at all.

• 02

  The six configuration layers (the Resonance thesis)

  You will work through the six configuration layers at the heart of the Resonance approach, which are task scoping, restricting which tools can be used, structured input and output, cost budgets and kill switches, an evaluation harness, and observability, and you will see exactly which failure mode each one is there to prevent.

• 03

  When to use an agent (and when not to)

  You will learn to decide when an agent is the right answer using a simple test, asking whether you would let a junior intern do this work with three checkpoints along the way, and weighing the cost against the human time it would save.

• 04

  Designing your first three agents

  You will use a specification template and walk through three common small and medium business patterns, ending up with agent specifications you can hand straight to a developer or a vendor.

• 05

  Human in the loop: what to gate, what to flow

  You will learn the discipline of designing where a human stays in the loop, telling apart the choices that could do real damage from the ones that could not, and building approval steps that genuinely matter rather than becoming rubber stamping.

This module is about getting AI to actually land across your team, starting with how to win over your most resistant colleague and how to train people without a formal training programme. You will learn to measure impact honestly, to keep momentum alive well past the first few months, and to recognise the moment to bring in a developer.

• 01

  The sceptic wedge

  You will understand why both mandates and outright bans tend to fail, and you will learn how to turn your most resistant colleague into a useful sceptic who actually strengthens the rollout.

• 02

  Training the team without a training program

  You will learn to bring your team up to speed without any training budget, using lunch and learns, a prompt of the week, and simple pairing patterns across a thirty day ramp.

• 03

  Measuring impact, properly

  You will reuse the metrics framework from the first module to measure impact honestly, telling real signals from vanity numbers and settling into a steady weekly rhythm.

• 04

  Sustaining momentum past month three

  You will learn why the early novelty tends to fade after the first few months and how to keep momentum alive, using a rotation pattern and a regular showcase to keep people engaged.

• 05

  When to call in a developer

  You will learn to recognise the boundary between what you can roll out without writing code and what genuinely needs engineering, and you will know how to brief a developer when you reach it.

Here we cover the Australian regulatory landscape that actually applies to you, including the Privacy Act 2024 reforms, the OAIC guidance on AI, the ten guardrails in the Voluntary AI Safety Standard, and your exposure around defamation and intellectual property. You will also work through data residency in practice, APRA CPS 230, and the state level frameworks, which is the part of this subject that almost nobody else bothers to teach.

• 01

  The Privacy Act 2024 reforms and OAIC AI guidance

  You will understand how liability attaches to personal information both going into and coming out of an AI system under the Australian Privacy Principles regime, and you will grasp the scale of the exposure, which can reach fifty million dollars in penalties.

• 02

  The Voluntary AI Safety Standard: ten guardrails

  You will walk through each of the ten guardrails in turn and see what putting each one into practice actually looks like for a business of around twenty five people.

• 03

  Defamation, IP, and copyright exposure

  You will learn how strict liability defamation works in Australia, how an employer can be held responsible for content an AI generates, and you will get an introductory grounding in intellectual property and copyright exposure.

• 04

  Data residency in practice

  You will explore APP 8 in depth alongside the My Health Records Act, learn which sectors face hard Australia only requirements, and understand the bring your own key pattern and where it fits.

• 05

  State frameworks, APRA, and sector regulators

  You will get to know the New South Wales, Victorian, and Queensland frameworks, an introduction to APRA CPS 230, and a clear sense of where the TGA, ASIC, and ACMA each come into play.

This module helps you keep AI healthy and growing over time, with evaluation writing that a non engineer can genuinely run, sensible monitoring, and a clear plan for incidents. You will also learn when to escalate to engineering and how to hire consultants without being taken for a ride.

• 01

  Eval authoring for non-engineers

  You will learn to write a weekly evaluation that your team will genuinely run, using rubric scoring and an approach to letting an AI act as judge that a non engineer can comfortably manage.

• 02

  Monitoring and incident response

  You will learn what is worth monitoring, what actually counts as an AI incident, and how to work through the question of whether something amounts to a notifiable data breach under the OAIC rules.

• 03

  When to escalate to a developer (deeper)

  You will learn to read the signs that you genuinely need engineering, use a briefing template to set the work up well, and interview a developer for AI work with confidence.

• 04

  Choosing consultants and vendors at this stage

  You will learn to spot the red flags in a consultancy pitch, understand realistic pricing benchmarks for AI work in Australia, and keep scope creep firmly in check.

• 05

  The next 12 months

  You will get a grounded picture of the model landscape ahead, including how model deprecation works and the question of consolidation, along with a sense of when it makes sense to move on to the course for developers.