Mike Cannon-Brookes Says SaaS Isn't Dead. He's Half Right.
By Libbie Ray | Founder, Connected Apps | Director, Connected Event Group
Mike Cannon-Brookes made headlines this week addressing the "SaaS is dead" narrative as Atlassian's share price plummeted close to 70%.
His take? AI enhances software, it doesn't replace it.
Mine? He's half right.
For enterprise, maybe traditional SaaS survives. Atlassian just posted $1.6 billion in quarterly revenue, 23% year-over-year growth, and five million users on its Rovo AI platform. The markets aren't buying it — but the enterprise customers are.
Here's the thing nobody in the "SaaS is dead" debate seems willing to say out loud: the conversation was never really about enterprise. Enterprise has always had the budgets, the IT teams, and the negotiating power to make bloated platforms work for them. They customise Salesforce with armies of consultants. They build Jira workflows so complex they need dedicated administrators.
The real story — the one that affects the vast majority of businesses — is what's happening to SMEs. And for SMEs? SaaS has been dying for years. AI is just making it impossible to ignore.
The Square-Peg Problem
Here's what I've watched happen over 15+ years running businesses in the events industry: Small and medium businesses forced to jam their square peg into a round hole. Paying thousands for bloated SaaS platforms built for enterprises. Trying to make workflows fit the software, instead of the other way around. Compromising on what they actually need because "that's just how the system works."
Sound familiar?
You sign up for a CRM that promises to "revolutionise your client relationships." Six months later, your team is using maybe 15% of its features, you're paying for seats nobody logs into, and the one workflow that actually matters to your business? It requires a "workaround" that involves three Zaps, a shared spreadsheet, and a prayer.
It's been broken for a long time. The difference now is that the tools to fix it are finally accessible.
The Numbers Behind the Shift
This isn't just my observation from the trenches. The data tells the same story.
Gartner forecasts that by this year, 75% of all new enterprise applications will be built using low-code or no-code platforms — up from less than 25% just a few years ago. SMEs are driving the fastest-growing segment of this shift, with an 18.7% compound annual growth rate in adoption. Organisations using these platforms report average annual savings of $187,000, with the majority seeing payback within 6–12 months.
And here's the stat that should make every SaaS vendor nervous: according to Gartner, 41% of employees are now "business technologists" — workers outside IT who are building their own solutions. Not because they want to play developer. Because the off-the-shelf options failed them.
The low-code/no-code market is projected to exceed $30 billion this year alone. By 2030, it's forecast to hit $101.7 billion. That's not a trend. That's a tectonic shift in who builds software and why.
Mike's Right About Enterprise. He's Wrong About Everything Else.
Let me be fair to Cannon-Brookes. When he says "software is dead" is a "ludicrous" statement, he's arguing from a position of strength. Atlassian's customers are predominantly mid-market to enterprise. They have compliance requirements, security needs, and organisational memory stored across years of Jira tickets and Confluence pages. For these businesses, switching costs are enormous and AI genuinely does enhance the existing platform.
His "input-constrained vs. output-constrained" framework actually makes sense for the enterprise world. Companies are currently limited by how much work they can do (input-constrained). AI lets them do more with the same resources. That's genuinely powerful.
But here's what that framework misses: SMEs aren't input-constrained. They're tool-constrained. They're constrained by software that doesn't fit, integrations that don't exist, and monthly subscriptions that add up to tens of thousands per year for features they'll never use.
The "SaaSpocalypse" that's sending shivers through Wall Street isn't really about AI replacing software. It's about AI enabling a fundamentally different approach to how smaller businesses get the technology they need.
What "Systemised Customisation" Actually Looks Like
The future isn't about choosing between generic SaaS or building everything from scratch. That's a false binary that serves the incumbent vendors.
The future is systemised customisation — getting exactly what you need, integrated with your workflows, without compromise.
Here's what this looks like in practice:
A 20-person events company that was paying $2,000/month across four different SaaS platforms — a CRM that didn't understand event production, a project management tool built for software teams, an invoicing system that couldn't handle deposit structures, and an email platform that didn't talk to any of them. We replaced all of it with a single custom system, built around how they actually work, for a one-time investment that paid for itself in under five months.
A conference with 2,000+ delegates and 200+ speakers drowning in email chains, USB sticks, and last-minute presentation panic. We built a custom speaker portal in two weeks that handled uploads, automated formatting, and real-time updates. It saved 40 hours of pre-event labour and resulted in zero presentation failures. No off-the-shelf event platform could have done exactly what was needed.
An equipment finance business struggling with lead tracking across multiple channels — referrals, website enquiries, dealer networks — all managed in disconnected spreadsheets. A custom CRM, purpose-built for their specific pipeline, now gives them visibility they never had. Not because Salesforce couldn't technically do it, but because the time, cost, and expertise to configure Salesforce for a team of 12 people was absurd.
These aren't edge cases. This is the everyday reality for millions of SMEs worldwide.
The Uncomfortable Truth About Per-Seat Pricing
There's another dimension to this that the enterprise-focused "SaaS isn't dead" crowd conveniently sidesteps.
The per-seat pricing model — the backbone of traditional SaaS revenue — is under genuine structural threat. Not from AI replacing employees (though that's happening in some categories), but from AI making it possible for fewer people to do more work without needing separate software seats for every function.
When Anthropic released its agentic desktop tool and Atlassian's own Rovo AI can handle tasks that previously required a human sitting in front of a Jira board, the logical question becomes: why am I paying per seat when the "seat" is increasingly occupied by an agent?
For enterprise, the answer is governance, compliance, and audit trails. Fair enough.
For SMEs? The answer is: you probably shouldn't be. You should be paying for outcomes, not seats. For systems built around your processes, not platforms you have to contort your business to fit.
Why This Is Personal
This is the entire reason I started Connected Apps.
After 15+ years in the event industry with Connected Event Group, I watched our sector struggle with this exact problem. Generic event apps that felt like glorified PDFs. CRM systems that didn't understand how event production actually works. Workflow tools that created more work than they solved.
The events industry is a brilliant microcosm for this conversation, because it's high-stakes, time-critical, and deeply human. You can't have a "workaround" when 2,000 people are in the room and the presentation system fails. You can't tell a client that their custom requirements are "not supported in this tier." You can't patch a broken workflow with a Zapier connection when the show starts in four hours.
Twenty years of that pressure taught me something that transfers directly to the broader business automation conversation: technology is only useful when it disappears. When the end user — whether that's a conference delegate, a sales team member, or a customer — forgets about the tech and just experiences a seamless process.
That's the standard generic SaaS consistently fails to meet for smaller businesses. And that's the standard custom, intelligent software can achieve.
What We're Actually Building
At Connected Apps, we build two things:
Interactive Apps & Activations — Bespoke digital experiences for events, activations, and retail. Not templates. Not "close enough." Exactly what the project needs. From custom presentation portals for Australia's largest cybersecurity conference to interactive experiences that drive real engagement.
Business Automation — Custom CRM and operational systems built around how businesses actually operate. Your workflows, your processes, your team. Not a platform you need a certification to configure.
Both are built on the same philosophy: start with the problem, not the tool. Understand the workflow before touching the tech. And deliver something that feels invisible — because it just works.
So, Is SaaS Dead?
It's the wrong question.
SaaS as a delivery model — software accessed via the cloud, maintained by someone else, paid for on subscription — that's not going anywhere. The mechanics are sound.
But SaaS as a philosophy — "here's our platform, fit your business into it, pay per seat, configure it yourself or hire a consultant" — for SMEs who've been compromising for too long, that version is dying. And it should.
The businesses that will thrive in the next era aren't the ones debating whether AI kills software. They're the ones asking a better question: "What would our business look like if every system was built around how we actually work?"
Custom, intelligent software that actually fits? That's just getting started.
The Positive AI Movement
This is what I mean when I talk about the Positive AI Movement. It's not about AI for AI's sake. It's not about replacing people with machines. It's about using artificial intelligence — combined with no-code tools, custom development, and genuine business understanding — to build technology that augments human capability.
65% of Australians believe AI creates more problems than it solves. I understand the scepticism. But the scepticism isn't about the technology. It's about the implementation. It's about AI being used to fire 150 people via pre-recorded video (as happened at Atlassian last year) versus AI being used to build a custom system that saves a small team 20 hours a week so they can focus on the work that actually matters.
The narrative needs to change. AI should be augmentation, not replacement. Ethics and transparency should be competitive advantages, not afterthoughts. And the $315 billion economic opportunity that AI represents for Australian businesses shouldn't be captured only by companies that can afford enterprise-level platforms and implementation teams.
It should be accessible to every SME willing to invest in building something that actually fits.
Libbie Ray is Director of Connected Event Group and Founder of Connected Apps (connectedapps.com.au). She speaks on AI automation, custom business applications, and the Positive AI Movement. She hosts The Conference Collective podcast for the Professional Conference Organisers Association and has presented at PCOA national conferences and EO Sydney.