Australian companies using Zendesk's AI tools are seeing measurable gains in customer support operations, from round-the-clock coverage to ticket triage in more than 60 languages. At the Zendesk Showcase event in Melbourne, support leaders from carsales, Catapult Sports, and Sealy of Australia described how automation has changed their service teams - not by replacing human agents, but by freeing them to focus on complex work.
Jon Aniano, Zendesk's senior vice-president and general manager of customer experience, argues that automating a large share of support resolutions doesn't remove a matching share of human effort from a service operation. Agents' time simply becomes less tightly rationed, and the interactions that do reach them get more of it.
Carsales: From suspicion to 24/7 support
Carsales has used Zendesk for almost a decade, said Rodney House, head of customer support and sales enablement. The platform now spans multiple business units, including consumer and dealer teams. Recently, the company adopted Zendesk's AI features for quality assurance, general call questions and answers, customer-facing bots, and next-best-action prompts for team members on calls.
At the start of the project, the team "had the feels but not the facts" about how AI could help, House said. They spent three weeks exploring the features before implementation. "They were initially suspicious of AI," he said, but the exploratory approach meant they could try different things and see what worked.
Zendesk's AI agents understand more than 95% of conversations and automatically resolve more than 80% of them. Contrary to expectations that this might reduce call and email volumes, the number of customer conversations rose - because AI gave carsales a 24/7 support team rather than one that works from 9am to 7pm.
"What we've seen is an increase in the amount of customers that we've been able to support because we're able to provide help when they need it, outside traditional business hours," House said. "For more straightforward enquiries, customers can get answers faster, while our team can spend more time helping customers with more complex needs."
Dealers generally prefer email to calling, but many of their queries are similar: "Why is my listing not up?", "Why is that photo doing this?", and often "Why is my ad being removed?" The usual answer to the last is that the ad contains branding elements, as listings are supposed to be "blind" rather than revealing the dealer's identity. Answers have been written for these low-tier questions to provide an instant response, and the AI has matched the team's performance on these tasks.
House said the knowledge spread through peer-to-peer learning, which has done more to improve the team's capability than any directive from the top. "The one thing that's come out loud and clear to me in all of this is AI is a great tool, but it makes people faster, not smarter," he said. "If you're going to go faster then you need to make sure that your checks and balances are keeping up, and that's one of the reasons why we went for evolution rather than revolution."
His advice for those starting with Zendesk or its AI tools: "Don't wait, do." Then, take the team on the journey - sit down with them and have a robust conversation about the plan. "AI is not the silver bullet that fixes all your problems, it is part of a toolkit that you should be using, and talking to your customers should be part of that as well."
Catapult Sports: Triage across time zones and languages
Catapult Sports provides wearable tracking devices and video analytics to teams in major football codes, baseball, and motorsport. "Sport doesn't wait," said Darcy Brown, product support manager. Users need accurate data and a quick response when problems arise.
Using Zendesk, Catapult has achieved a 100% customer satisfaction score, with a median first response time of six minutes and a median resolution time of one hour. Support is delivered from Melbourne, Leeds and Boston to provide round-the-clock coverage, said product support team lead Andrea Juniper, with smaller teams in Mexico, China and Japan.
The company has grown organically and through acquisitions, so the volume of support requests has outstripped the expansion of the support team, Brown said. Catapult uses Zendesk AI agents to triage tickets, solving problems where they can - currently 36% of cases - and passing more complex issues to human agents. All of the context collected by the AI agents is delivered to the human when an issue is escalated.
Zendesk's AI translation capabilities "changed our life", said Juniper, as Catapult's customers speak more than 60 languages compared with the four spoken by support team members. The feature processes tickets and provides voiceovers for help content in more languages.
Using these capabilities has cut first response time by 52%, resolution time by 25% and handling time by 21%, Juniper said. Brown identified several business outcomes, including a reduced cost to serve a growing customer base, and the ability to provide elite teams with an elite experience that is fast where it can be and personal where it should be.
For those planning to adopt Zendesk's AI features, he suggested starting automation with high-volume but low-risk requests, involving representatives of the wider business at an early stage, and continually proving the return on investment at every step.
Sealy of Australia: Fixing the foundations first
During the Covid-19 pandemic, mattress maker Sealy of Australia saw enquiry volumes rise to the point where two weeks were passing before customers got a reply, said Michael Chen, head of digital and customer experience. Hiring more staff didn't make sense for what was expected to be a short-term problem, so the company started a journey aimed at futureproofing this part of its operations for the next 20 years.
Chen's approach was to address the people and culture aspects first, then the processes, and only then the technology. His reasoning: AI amplifies whatever you have, so if you don't get the foundations right, it will just help you make the same mistakes more quickly.
Now, Zendesk Copilot reduces repetitive tasks, and AI agents handle common questions and escalate enquiries when necessary. People are still needed to resolve complex situations, and sometimes to read between the lines and understand what a customer means but is not saying.
On return on investment, Chen said saving two to three minutes per call quickly adds up and was enough to cover the cost, so any other uses provide a bonus.
Why this matters for customer support professionals
Three patterns emerge from these accounts that apply to support teams beyond the Australian market. First, AI adoption works best when it starts with experimentation and peer-to-peer learning rather than top-down mandates - carsales spent three weeks testing features before committing. Second, automation changes the shape of the workday: teams that adopt AI agents effectively gain a 24/7 presence without expanding headcount, while human agents spend their time on complex cases that need judgment. Third, the economics work at a practical level - Chen's calculation that saving two to three minutes per call covers the cost of the tool is a benchmark any support leader can apply to their own operation.
For support professionals looking to build these skills, resources like AI for Customer Support and AI for Call Center Supervisors offer practical training aligned with the workflows described here - ticket triage, escalation handling, and managing AI-assisted teams.
Your membership also unlocks:
