Defensive Moats in the AI Era: Why Vertical Micro-SaaS Is Hard to Displace
The “SaaS Is Dead” Narrative and the Reality on the Ground
When large language models allow anyone to assemble a software prototype in hours, the question echoing among Australian founders is: is there still room for small SaaS? Matthew Attanasio, founder of Sign Customiser, hears that question every week. For him, the answer lies in one word: specialization.
Sign Customiser is a design and quoting tool for the neon sign industry. It sounds narrow. Yet in 2025, the platform facilitated nearly $17 million in merchant sales from more than 800 clients across 50 countries . When asked whether an in-house team at a large company could “vibe code” a similar tool with AI assistance, Attanasio did not deny it. He simply cited a number: reaching the same level of refinement could cost hundreds of thousands of dollars in AI compute tokens. And by the time they finish, he is already five steps ahead .
Feedback as Growth Fuel
The advantage of vertical micro-SaaS is not in the code. It is in the accumulation of domain knowledge that no AI model can replicate without the same data. Attanasio describes it as a feedback loop from hundreds of merchants worldwide guiding him on what will help them handle more sales . This information is not available on the public internet. It flows only through commercial relationships built over years.
A similar pattern appears in other segments. Toastie, a health-tracking app built by two Google engineers as a side project, refuses to label itself an AI company despite using the technology internally . Its founders—Anna Zhou and Vivian Shen—explicitly stated that general AI models like ChatGPT and Gemini “aren’t good enough when the stakes are this high” . They built a custom logic layer to understand symptom patterns, food, and lab reports. The decision not to rely on generic AI is a form of self-defense against commoditization.
An Ecosystem Beginning to Value Depth
Mike Cannon-Brookes of Atlassian once stated that the idea “software as a category is dead” is absurd . Niki Scevak of Blackbird added that the winners in the AI era are not those doing vibe-coded custom solutions, but those who understand that large companies have historically failed to build their own software . These statements are not defensive rhetoric. They reflect market observation that the maintenance cost of internal software far exceeds the subscription cost of specialized SaaS.
For Australian micro-SaaS, the implication is clear: the best defense is not invisible technology, but depth of relationship with a market niche. AI lowers the barrier to building a product. It does not lower the barrier to understanding a problem worth solving. That is the dividing line between businesses that endure and those that get swept away.
