After eight years running the startup accelerator Startmate – which launches its latest cohort before a sold out crowd in Sydney on Thursday night – Michael Batko decided so solve a problem hundreds of founders asked him about: AI.
Every founder asked him the same question about artificial intelligence: “How do we actually use AI – in our business, right now?” Batko said. Nobody had an answer. So he decided to create it.
“I looked around at who was answering that question properly and couldn’t find anyone,” he said.
“Consulting firms gave them strategy decks. Agencies gave them discovery phases. Nobody just built the thing. So I left to go and provide what was missing.”
Batko teamed up with Finlay Ekins, 22, to cofound Hourglass AI – an AI implementation firm that builds working systems inside Australian businesses on a fixed-fee basis.
Get the best of Startup Daily straight to your inbox
Want to know the latest in startup news? Subscribe to our daily news and analysis coverage on what’s happening to ANZ startups, investors and the broader ecosystem. And best of all, it’s FREE!
By continuing, you agree to our Terms & Conditions and Privacy Policy.
It kicked off in stealth mode in February and two months on, launched this week.
The execution gap
The problem they’re solving is the execution gap in implementing AI.
Batko points to analysis Anthropic’s own research team published last month pointing to yawning gap between what AI systems can do and what businesses have actually built.
He’d already spotted it and says it was the reason he walked away from one of most influential jobs in Australia’s startup ecosystem.
“Everyone feels behind on AI, and they are,” said Batko.
“But it’s not because they don’t understand it. It’s because they’re too busy to actually build it. The operators who’ll win the next decade aren’t the ones reading about AI. They’re the ones with working systems in production right now.”
The startup veteran said Hourglass AI’s offer is the opposite of the industry default: fixed scope, fixed fee, live system in weeks. No strategy documents. No discovery phases. No slides.
“Every Australian executive reading this has sat through the same conversation. A board meeting about AI strategy. A LinkedIn post warning them they’re behind. A newsletter from a consultant promising a roadmap,” Batko observes.
“Almost none of them have shipped a working AI system inside their business.”
Already running
Hourglass AI’s early client roster spans retail, property, ecommerce and wellness, as part of a deliberately broad cross-section, selected to stress-test the mode.
They include Schweigen, a multi-billion-dollar appliance retailer, which now runs an AI customer support system built by Hourglass, live on the company’s own site; Cadence Property Group, which is being transformed end-to-end, with 12 of 35 staff audited, an AI knowledge brain under construction, the whole company being brought onto Claude, with custom agents. The others are The Skincare Company, which is integrating AI across internal operations to streamline the business while keeping its customer-facing experience organic and personal; and Sense of Self, a Melbourne wellness destination, which is having core operations streamlined with Hourglass-built AI systems.
Batko explains that Hourglass has a deliberate three-phase plan.
Phase one is what he calls “getting paid to learn”: embedding inside client businesses, shipping real systems, and understanding what actually works at the coalface rather than a lab.
Phase two takes that hard-won knowledge to scale it across thousands of businesses.
Phase three is the long game: building the best AI-native company in Australia from the ground up.
“Australians are world-class at execution when you give them the room,” said Batko.
“Right now we’re in phase one – learning the craft properly inside real businesses, with real problems, shipping real systems. Everything we build from here gets built on that foundation. We’re not here to be a boutique agency forever. We’re here to do something significant.”
Watch and learn
The starrtup is also launching two public initiatives designed to close the execution gap at scale.
The first is an upskilling programme for operators who want to build AI systems themselves.
The second is a weekly live-build stream – Ekins and Batko publicly shipping AI systems in real time, so anyone watching can see exactly how the work is done, start to finish.
“Most people don’t learn AI from a course. They learn it by watching someone build one real thing from nothing,” Batko said.
“We’re going to be the builders who show their working.”
His young cofounder, Finlay Ekins said AI is so new that anyone willing to put in the effort can become an expert, regardless of age.
“So I skipped the grad job and rewrote the playbook from scratch, including the one that said how to run a company,” he said.
“The businesses willing to do the same are going to end up light-years ahead of the ones sticking their heads in the sand.”
AI builder jobs
The next step for the Hourglass AI duo is hiring for two AI builder roles. They want to find young Australians who want to spend their careers actually shipping AI systems inside real businesses.
The hiring process mirrors the firm’s operating model: applicants receive five real meeting transcripts and have seven days to build the best AI-powered experience for follow-ups, summaries and next steps..
“If you’re young and you want to build AI systems – not read about them, not advise on them, build them – this is the job,” Batko said.
“Australia produces world-class builders. We want to find them and give them the best possible place to do the best work of their careers.”
To find out more, visit thehourglass.ai.
