Australia is ten years behind on enterprise AI adoption

The second episode of Beyond the Bubble moves the conversation out of San Francisco and into a market that, by the guest’s own admission, is still treating ChatGPT as a chat interface. Muzamil sits down with Jenny Vo, founder and CEO of Hera Digital, to understand what AI deployment actually looks like inside corporate Australia, and why the gap between the narrative and the reality is wider than the headlines suggest.

A bubble sitting on top of a foundational shift

Muzamil opens with the show’s thesis. He is “100% in a financial bubble around AI,” but believes the underlying technology is foundational enough to reshape the world. The question he keeps returning to is what survives once the financing narrative cools. Jenny is a useful test case. She left corporate life and built Hera Digital, now almost two years old, six headcount, with roughly 80% of revenue from North America and 20% from Australia. The Australian split is small on purpose. “The Australian market is still very, very reluctant,” she says, and she traces that reluctance to a missing education layer rather than a missing willingness to buy.

Predictability is the hard part, not AI

Muzamil pushes on the tagline on her site, “Build predictable sales systems with AI”, noting that sales systems are not new. Jenny pushes back. “I actually like to challenge you on that perspective, because the AI piece now has actually become overused and oversaturated. It’s more the predictability side that is actually the hard component.”

In her framing, AI works best as a data analytics layer on top of the pipeline, not as a magic wand. Hera tends to start engagements with what she calls an ICP magnification process, rebuilding the RevOps engine from ideal customer profile through to close, then layering agentic AI on the parts humans consistently fail at. The two failure modes she sees again and again are a poor ICP and poor follow-up. Usually both.

The 69% number

The sharpest stat in the conversation comes from joint work Hera did with a sales coach. “69% of the pipeline dies at conversion due to the lack of follow-up,” Jenny says. Her explanation is psychological, not technical. Salespeople feel monotonous reaching out for the third or fourth time, they do not want to bother the prospect, the gap widens, urgency dies. This is the seam where Hera inserts automation, but with a deliberate constraint: the follow-up has to read as human. “Everyone can read AI slop,” she says. Her practical tip for operators sending automated outbound is to purposely include small mistakes, because polished output is now a tell.

Why regulated Australia is ten years behind

Muzamil presses on why adoption is so slow. Jenny is blunt. In Western Australia, where Hera is based, the economy is mining-heavy, and growth for those clients does not mean more revenue, it means more time back. In finance and other regulated sectors, the brake is compliance. “All the sectors that are heavily regulated are still, I would say, comfortably ten years behind in adoption,” she says, because almost no one is doing the work of AI governance and safe deployment inside those compliance walls. Founders she has spoken to in finance simply say they do not know where to start.

She also confirms this is not AI-specific. A recent guest on her own podcast, a Cisco veteran, told her the same reluctance hit Australia when the worldwide web arrived. Her practical conclusion for founders listening: if you want to enter Australia, do it through partnerships with operators who already hold the local IP and the behavioural knowledge. “Unless you have that IP, I don’t see the business lasting very long in the Australian market.”

The post-AI fantasy and the human correction

Muzamil widens the lens. He points out that much of the global south, including his own context in Pakistan, has resisted technology for forty years not because of cost but because of human aversion to change. He asks Jenny whether the X timeline, where every white-collar job is automated in eighteen months, matches reality.

Her answer is one of the most useful moments in the episode. “I don’t think there is going to be post AI.” AI will keep evolving and being deployed, but a business that deploys AI specifically to replace humans is, in her view, making a bad call. “It cannot feel, it cannot experience, it cannot express.” She describes an imagined A/B test where channel A is pure AI outreach and channel B is human, and predicts B wins on response rate because the market is overcorrecting against AI-generated contact. The equilibrium, she argues, will require humans back in the loop, partly to operate AI safely.

From novelty to outcomes

Muzamil names what he is seeing across the market. People are tired of AI being pushed into their face. He remembers a Logitech AI mouse that turned out to be a button that opened a chatbot. His read is that the FOMO framing of 2024 is dead, and the conversation is shifting to outcomes. “It’s no longer about AI. It’s about outcomes,” he says. AI has to be sold the way mobile apps were sold, not as a technology but as a way to sell pizzas, or close deals, or save time.

Jenny agrees and adds a counterintuitive data point from inside her own company. On a recent Monday, Hera did a full silo-by-silo audit of where AI could be inserted into the business. “Six times out of ten, I still prefer to human,” she says. Her partner, she notes, would rather wait on hold than talk to an AI agent. She is not arguing against deployment. She is arguing the cost of removing humans is not only financial. Running a company alone, with only AI agents instead of colleagues, she says candidly, would be unbearable.

The Noodle Seed partnership and the chicken-and-egg problem

The conversation turns to how Hera gets traction in a reluctant market, and Jenny points to her partnership with Noodle Seed and its work on custom ChatGPT apps. Her logic is commercial. Building that IP in-house would be high effort and low impact for Hera. Plugging into a partner who has already built it lowers the barrier of entry for her clients and removes the risk. “People are not willing to pay for something they don’t understand,” she says, and AI carries an unusually heavy version of that problem.

Muzamil draws the parallel that interests him most. Noodle Seed sits in roughly the position WordPress occupied around 2006, or mobile app builders occupied a few years later: the next layer of digital presence, this time built for a world where 800 million people start search inside a chatbot. The catch is the chicken-and-egg problem. The ecosystem needs enough early adopters to justify the build, but early adopters need the ecosystem to justify the spend. Jenny’s answer is that the businesses that survive this gap are the ones built around the end user rather than an ARR target. She points to Noodle Seed itself, which pivoted 180 degrees away from vibe coding into its current shape, as evidence that prioritising the user is what creates a defensible position.

SaaS is not dead, it just has to pivot

Muzamil closes the substantive part of the conversation on the “SaaS is dead” narrative that ran through 2024 and 2025. Jenny rejects it cleanly. “I don’t believe that SaaS is dead.” She describes what she calls the “crackhead era” of opening Claude Code for the first time, learning what is possible, and then realising six hours of custom building at a $100 hourly value is $600 you could have spent talking to customers. The commercial logic, she argues, still favours paying twenty dollars a month for software someone else maintains.

Where she sees real expansion is in subverticals that were previously too small or too unglamorous to justify a SaaS build. Muzamil’s read, which she agrees with, is that as the cost of producing software collapses, hundreds of new use cases open up in industries that were historically untouchable. She mentions meeting founders who have digitised the charity sector specifically to give donors visibility into where their money goes, which she frames as a SaaS product wearing an NGO label.

What Hera is actually for

Muzamil asks Jenny what the ideal version of Hera Digital looks like. Her answer pulls the episode back to the human thread that has run through it. She wants to do more work with mission-led and purpose-driven organisations. Her reasoning is that dentists are good at being dentists and bakers are good at being bakers, and forcing those people to also be business owners dilutes the craft that made them worth serving in the first place. If Hera can take the operational pain off those operators, she says, “I’ll die happy.”

It is a fitting close for a show whose premise is that the durable AI businesses will be the ones still standing when the financing narrative breaks. Jenny Vo’s argument, across roughly an hour with Muzamil, is that the durability comes from the same place it always has: a clear end user, a real outcome, and a willingness to sit in the slow markets long enough to educate them.