181 lines
13 KiB
Markdown
181 lines
13 KiB
Markdown
Presenter Script: The AI Advantage
|
||
|
||
Title: The AI Advantage: Rethinking 'Build vs. Buy' in Government IT
|
||
|
||
Format: 30-Minute Presentation to 2026 Government Digital Graduates
|
||
|
||
Presentation Framework: Reveal.js (HTML5/CSS3/JS)
|
||
|
||
Presentation Setup & Key Controls
|
||
|
||
Before presenting, make sure you are comfortable with these built-in Reveal.js controls:
|
||
|
||
Open Speaker Notes (The Secret Weapon): Press **S** on your keyboard. This opens a separate presenter window showing:
|
||
- A timer/chronometer to keep track of your 30-minute block.
|
||
- The current slide and the next slide preview.
|
||
- The complete speaker notes script.
|
||
|
||
Full-screen Mode: Press **F** to go full-screen in your browser.
|
||
|
||
Help Overlay: Press **?** to toggle the shortcut help menu.
|
||
|
||
Pacing & Timing Guide
|
||
|
||
- **Slides 1–3 (The Irony, The Bill & The Capture Pie):** 8 minutes (Setting the scene, the Australian context, and the systemic wedges of vendor capture).
|
||
- **Slides 4–6 (The Disasters & The Outage):** 8 minutes (BOM blowout, the CrowdStrike outage, concentration risk & SLAs).
|
||
- **Slides 7–9 (Global Proof & The AI Shift):** 8 minutes (Germany/France migrations, why AI changes the math, keeping tech budgets in QLD).
|
||
- **Slides 10–12 (The Choice, Playbook & Closing Q&A):** 6 minutes (Sovereign building, how to start tomorrow, and final discussion).
|
||
|
||
---
|
||
|
||
# Slide-by-Slide Script
|
||
|
||
### Slide 1: Title Slide
|
||
**Slide Elements:** Clean, minimalist titles, subtle dark background, glowing orb animations.
|
||
**Transition:** Fade out.
|
||
**Speaker Script:**
|
||
"Good afternoon, everyone. Thanks for coming to SparkLabs. I want to start by acknowledging the massive, glaring irony of this presentation.
|
||
|
||
I am about to give a talk on why government needs to stop relying on mega-vendors. And I am delivering it to you right now over a Microsoft Teams call, while most of you are watching on Windows laptops, probably taking notes in Microsoft OneNote.
|
||
|
||
We are completely captured. Today, we are going to talk about how much that capture costs the Australian taxpayer, and how you—as the next generation of IT leaders—can actually do something about it."
|
||
|
||
---
|
||
|
||
### Slide 2: The Megavendor Rent (The Australian Context)
|
||
**Slide Elements:** Large text showing "VSA6" and "$1.6 Billion".
|
||
**Transition:** Slide in, fade out.
|
||
**Speaker Script:**
|
||
"Let's talk about the bill. In July 2026—literally this month—the federal government's new Volume Sourcing Agreement, or VSA6, kicks in. This locks the Commonwealth, and by extension sets the standard for states like Queensland, into the Microsoft ecosystem until 2031.
|
||
|
||
In the previous agreements, the government proudly announced they secured $1.6 billion in discounts across single seller arrangements. I want you to think about that math. That $1.6 billion discount isn't just one contract; it's the combined discount across whole-of-government agreements with Microsoft, SAP, Oracle, and others over five years.
|
||
|
||
If the discount alone is 1.6 billion dollars, what is the actual baseline cost? It is hidden across dozens of separate agency budgets. But we know the scale: across all vendors, federal IT spend topped 5.8 billion dollars last year. We aren't simply buying software anymore; we are paying a permanent rent on our own digital infrastructure. And when the landlord raising the rates knows you can't afford the moving truck, they hold all the cards."
|
||
|
||
---
|
||
|
||
### Slide 3: The Risk Aversion Paradox (The Vendor Capture Pie)
|
||
**Slide Elements:** A grid card showing QLD Health disaster metrics, and an SVG "Vendor Capture Pie" showing the 4 wedges: Risk Aversion, Capability Gap, Procurement Rules, and Migration Trap.
|
||
**Transition:** Slide in, fade out.
|
||
**Speaker Script:**
|
||
"So why do incredibly smart executives keep signing these deals? It's often blamed entirely on risk aversion—the old saying that 'nobody ever got fired for buying Microsoft or IBM.'
|
||
|
||
But risk aversion is just one wedge of a larger systemic pie. The QLD Health Payroll disaster is prime evidence: we contracted IBM for 6.2 million dollars, thinking we outsourced the project risk. Instead, it blew out to 1.25 billion dollars, caused 35,000 payroll errors, and took a decade to recover from. We retained all the risk, but outsourced the capability.
|
||
|
||
The other wedges of this capture pie are just as powerful: the Capability Gap, because decades of outsourcing have left us without internal engineering talent; Procurement Rules that are so complex only multinational giants can navigate them; and the Migration Trap, where our systems are so deeply entangled with proprietary models that the cost of moving feels impossible."
|
||
|
||
---
|
||
|
||
### Slide 4: Not Just a Local Problem (Bureau of Meteorology Redesign)
|
||
**Slide Elements:** Metrics showing $4.5M original budget vs $96.5M blowout.
|
||
**Transition:** Slide in, fade out.
|
||
**Speaker Script:**
|
||
"If you think that's just an old Queensland problem, look at the federal level. In late 2025, it was exposed that the Bureau of Meteorology website redesign—which originally had a 4.5 million dollar budget—blew out to 96.5 million dollars.
|
||
|
||
And the result? Australians couldn't even read the weather map properly. It infuriated farmers and regional citizens to the point where portions of the site had to be rolled back to their original views. It sailed through without proper oversight because the front-end redesign was anchored to an 80 million dollar external technical rebuild.
|
||
|
||
We keep throwing hundreds of millions at external vendors for basic digital touchpoints, and we end up with products that are functionally worse for the citizens who rely on them."
|
||
|
||
---
|
||
|
||
### Slide 5: The CrowdStrike Outage (Audience Poll)
|
||
**Slide Elements:** CrowdStrike Outage title. Left column: poll details. Right column: SVG radiating node diagram showing global impact and key stats (8.5M devices, 5,000+ flights, $5.4B Fortune 500 losses).
|
||
**Transition:** Zoom in, fade out.
|
||
**Speaker Script:**
|
||
"Before we move on, I want to pause and hear from you. Where were you on July 19th, 2024? Raise your hand on Teams or drop a quick note in the chat if you were personally affected by this. Were you stuck at an airport? Trying to pay for groceries? Locked out of a critical government system?
|
||
|
||
[PAUSE FOR 10-15 SECONDS TO LET CHAT POPULATE / READ OUT 1 OR 2 REPLIES]
|
||
|
||
For anyone who somehow missed it, or managed to block it out of their memory: that was the day of the global CrowdStrike outage. A single, routine configuration update pushed by one security vendor crashed eight and a half million Windows machines worldwide. It grounded flights, halted hospital surgeries, and crippled banking and government infrastructure.
|
||
|
||
Look at the numbers on the screen. 8.5 million machines in a Blue Screen of Death loop. Over 5,000 flights cancelled on day one. An estimated 5.4 billion dollars in direct losses for Fortune 500 companies alone, and over 1 billion dollars in damage to the Australian economy.
|
||
|
||
Now, raise your hand if you think CrowdStrike fully compensated the Australian economy for that damage.
|
||
|
||
[PAUSE]
|
||
|
||
Of course they didn't. Standard software contracts cap liability at the fees paid. We absorb the damage, while they send service credits."
|
||
|
||
---
|
||
|
||
### Slide 6: Too Big To Fail (Cloud Concentration & The SLA Myth)
|
||
**Slide Elements:** Comparison cards of CrowdStrike Reality vs SLA Illusion.
|
||
**Transition:** Slide in, fade out.
|
||
**Speaker Script:**
|
||
"When a major vendor goes down, they are essentially 'too big to fail.' What we used to call vendor lock-in has evolved into Cloud Concentration Risk. Three vendors control two-thirds of the world's infrastructure. When they go down, the state stops.
|
||
|
||
And the moment you suggest building locally, someone will say, 'But we need a vendor SLA for protection.'
|
||
|
||
Let's be realistic about what an SLA actually is. If an Azure or AWS outage costs a department $150,000 in lost productivity, standard vendor SLAs typically cap liability or offer a 10% credit on your next monthly bill. That’s maybe a $5,000 credit for a massive operational disaster. The SLA doesn't protect our operations; it protects the vendor's liability. We take all the actual risk."
|
||
|
||
---
|
||
|
||
### Slide 7: Real Sovereignty (The Global Rebellion)
|
||
**Slide Elements:** Cards for Schleswig-Holstein (Germany) and GendBuntu (France) migrations.
|
||
**Transition:** Slide in, fade out.
|
||
**Speaker Script:**
|
||
"If you think ditching these megavendors is a pipe dream, look at Europe. Other governments are waking up.
|
||
|
||
Right now, the German state of Schleswig-Holstein is actively migrating tens of thousands of public servants off Windows and Microsoft Office, entirely over to Linux and LibreOffice. Why? Their minister explicitly stated it was about Digital Sovereignty. They realized that handing their government's operational capability and data architecture to a foreign corporation was a national security risk.
|
||
|
||
France's national police force did the exact same thing years ago, saving millions and taking back complete control of their digital borders. It is entirely possible."
|
||
|
||
---
|
||
|
||
### Slide 8: The AI Multiplier (Why Now?)
|
||
**Slide Elements:** Graph showing historical cost of custom dev vs modern AI-assisted dev.
|
||
**Transition:** Slide in, fade out.
|
||
**Speaker Script:**
|
||
"The traditional pushback in Queensland has always been: 'We don't have the armies of developers needed to build and maintain our own open-source solutions.' And five years ago, that was a completely fair argument.
|
||
|
||
But that argument is dead. AI has completely changed the economics of software development.
|
||
|
||
A small, cross-functional internal team using AI coding assistants and modern lightweight stacks—like React, Vite, and TypeScript—can now scaffold, build, and secure custom open-source infrastructure in weeks. Building sovereign software is now demonstrably cheaper and faster than adapting our workflows to fit a vendor's rigid template."
|
||
|
||
---
|
||
|
||
### Slide 9: The Local Economy (Where Does the Money Go?)
|
||
**Slide Elements:** Visual flow comparing offshore transfer vs Queensland investment.
|
||
**Transition:** Slide in, fade out.
|
||
**Speaker Script:**
|
||
"This isn't just about code; it is a macroeconomic choice. The Queensland Government's ICT spend routinely runs into the billions.
|
||
|
||
When we sign a massive SaaS contract, a huge portion of those taxpayer dollars goes straight to a server farm in Seattle or Silicon Valley. If we take even a fraction of that licensing budget and redirect it to building open-source sovereign capability, where does that money go?
|
||
|
||
It stays in Queensland. It pays for local graduate jobs—jobs for people like you. It funds local Brisbane tech startups who we can hire to support and integrate our open-source tools. Public tech budgets should be used to build the local economy."
|
||
|
||
---
|
||
|
||
### Slide 10: The Choice for 2026 Grads
|
||
**Slide Elements:** Two choice cards: License Administrator vs Sovereign Builder.
|
||
**Transition:** Slide in, fade out.
|
||
**Speaker Script:**
|
||
"You are the 2026 digital graduates. In five years, you are going to be the lead enterprise architects, the senior BAs, and the design directors for the Queensland government.
|
||
|
||
You have a choice regarding what your career looks like. You can be a professional license administrator, managing vendor lock-in.
|
||
|
||
Or, you can be builders. You can champion open standards, use AI to create incredible local systems, and actually own the infrastructure that runs this state. When something breaks, you won't be submitting a ticket to a vendor in another timezone; you'll be fixing it because you own it."
|
||
|
||
---
|
||
|
||
### Slide 11: The Playbook
|
||
**Slide Elements:** 4-step playbook: 1. Demand Plan B. 2. Find a clunky process. 3. Build prototype. 4. Show, don't tell.
|
||
**Transition:** Slide in, fade out.
|
||
**Speaker Script:**
|
||
"How do we actually start? You don't walk into work tomorrow and try to aggressively unplug a critical legacy system. You start at the project level.
|
||
|
||
For your next assignment, when the team defaults to 'Let's put it on Azure' or 'Let's just buy a SaaS product,' push back politely. Raise your hand and request an open-source viability test. At the very least, ask what our 'Plan B' is if we need to migrate off the vendor later.
|
||
|
||
Then, find a terrible, clunky manual process in your department. Use an AI assistant and a lightweight framework to scaffold a clean, secure, local internal web app to fix it. Show leadership a working tool built in three weeks with zero licensing fees. The best way to win the argument is simply to build something better."
|
||
|
||
---
|
||
|
||
### Slide 12: Conclusion
|
||
**Slide Elements:** Large text: "Public Funds. Public Infrastructure."
|
||
**Transition:** Fade in.
|
||
**Speaker Script:**
|
||
"Digital infrastructure is exactly like physical infrastructure. We wouldn't let a private foreign company own the M1 motorway and charge Queenslanders a toll every time they drive to work. We shouldn't let them own the digital highways our citizens rely on either.
|
||
|
||
Public funds should build public infrastructure. Let's start building it.
|
||
|
||
Thanks for listening. We have about eight minutes left, and I'd love to open the floor up. To kick us off: what proprietary tools are driving you crazy right now, and what could we realistically replace them with?"
|