End of financial year hits Hnry from three directions at once.
Every year, Australian end of financial year hits Hnry from three directions at once. Support volume climbs as customers try to understand their tax position. Onboarding surges, because everyone waits for the new financial year to sign up, which pushes first payments, ID checks, and AML checks up alongside it. And on top of that, it’s also when the team is working through the actual mechanics of reviewing and lodging thousands of income tax returns.
Alex MacAvoy, Head of Customer Operations, is blunter about the scale of it. It is, she says, the most challenging week of the year.
Simple questions go to the agent, judgement calls go to the team.
Hnry’s view going in was that not every one of those conversations needed a person. A lot of questions have one factual answer, and when that’s the case, most customers would rather get it straight away than wait in a queue for someone to type it out. So the plan was to let the agent take those questions, the simple, repeatable ones, and free up the team for the conversations that actually need a person: where the situation’s complex enough that it needs real judgement, or where a customer just needs to talk to someone rather than be given an answer.
That’s also what gives the team more time to get ahead of things instead of just reacting. A simple question gets answered on the spot, any time of day, no queue. And the time saved lets the team spot a pattern in someone’s account before it becomes a problem, or notice an expense a customer hasn’t claimed and let them know, rather than waiting for the customer to bring it up.
Hnry deployed Lorikeet on Australian tax first, the hardest jurisdiction it has.
Australian tax rules carry enough nuance that a correct answer often depends on information only the customer holds, and Hnry relies on them to provide it accurately. As Staci puts it, between the rules, the nuances, and how quickly customers can get confused and frustrated, it adds up to the hardest support environment the company has. It is also the jurisdiction Hnry chose to deploy Lorikeet on first.
Australians grow up expecting a tax return process that runs through an accountant and ends in a refund, self-employed or not. Many are counting on that refund, or at least waiting on it, which adds urgency to the conversation. New Zealanders don’t carry the same expectations, so when an Australian customer’s experience doesn’t match what they’re used to, the frustration shows up faster.
That mismatch between expectation and outcome shows up clearly in one of the tickets that costs the team the most time: a customer wanting to understand why they’ve been taxed at the rate they have. Answering it means walking through how their self-employed income adds up, accounting for things like other income sources and deductions, and explaining how all of that comes together to land them on the specific rate that applies to them. The question comes up often enough, and the answer follows a consistent enough shape, that Lorikeet now handles it as a structured FAQ-style response rather than passing it to a human.
From a standing start to around 70% automation in the peak week.
Hnry’s agent went live in mid-May 2026, from a baseline of zero automation. In the first month it handled 17,000+ support conversations.
By the peak week of Australian end of financial year, Lorikeet was automating around 70% of them, and independently resolving around 28% end to end with no human involved at any point. The customer’s issue was closed out by the agent alone.
Ticket volume climbed through late June and peaked at the turn of the financial year. Automation climbed with it, from roughly 58% across the full period to around 70% in the week the load was heaviest.
New Zealand and the UK are live, with a wider knowledge base next.
Lorikeet is already live in New Zealand and the UK as well as Australia. The current focus is broadening the knowledge base so it can answer a wider range of questions in each market, and looking at ways to integrate it more deeply into how customers manage their accounts.
