Young Kiwis in AI: April Meetup

The Young Kiwis in AI meetups are a chance for young people working with AI to share what they are building, what they have learned, and what they would do differently. The April 2026 session featured two member spotlights: Courtney King from Stitch Systems on marketing and automation, and Donncha Hand from BlueHorizon AI on automating the insurance industry.

Courtney King: marketing meets automation at Stitch Systems

Courtney is a marketing and automation consultant who recently made her business official as Stitch Systems. The name is part dog, part method: her dog Lilo is her world, so the business became Lilo and Stitch, with Stitch reflecting how she stitches together her marketing background and her new AI and automation skills.

She is a BCom marketing and management graduate who has worked in-house and freelance. Her curiosity sparked at Robert Walters, where a culture of researching and reporting back led her to present to her team on how to use AI and set up prompts properly. Since then she has been sharing what she learns on TikTok and YouTube, which is how she first met Blake, and she now heads up the automation programme at Human Resources New Zealand alongside her own client work.

The coolest thing about where AI is now, for Courtney, is accessibility. With a marketing background and plenty of hands-on time with clunky websites, she appreciates that you can now stand up a site with Claude Code doing most of the work. There is plenty of slop that comes with that, but building a decent website used to take a ridiculous amount of time or cost thousands of dollars, and that barrier has largely fallen away.

Her work reflects a careful, honest take on AI in marketing. She is not 100% sold on it, sharing the common concern about AI slop and the loss of authentic human creativity, especially in visual content, but argues that for the right business, set up properly, it is a no-brainer. For the HR NZ marketing team she uses Claude Code to build scripts they trigger from Slack, plus weekly automations that feed them content inputs. For a small trades business, uploading a photo to Google Drive generates a social post, which solves their real problem: a camera roll full of before-and-afters they never get around to posting.

Her biggest current project is digitising HR NZ's professional accreditation process, today a heavily manual flow of Word docs emailed to an accreditation manager, with a charter assistant bot to walk people through it and point them to relevant courses and resources.

On lessons, her standout failure was not knowing which tool to use when. She once spent ages debugging a website in plain Claude before realising she should have been in Claude Code, where she fixed it almost immediately. Her advice to her past and current self is to stop consuming and start doing, and to focus on one thing at a time rather than juggling a million projects.

Her tips: do an honest audit of your own expertise and lean into the unique perspective that will differentiate you as AI becomes more prevalent; learn Claude Code, where she rates creator Nick Saraev's tutorials; and talk about what you are doing, because most of her leads have come through word of mouth and you never know where the next one will come from.

Donncha Hand: automating insurance at BlueHorizon AI

Donncha Hand, who goes by Don, is the co-founder of BlueHorizon AI, an Auckland agency he runs with his co-founder Caique Rivero. They have been going properly since around early 2025, roughly a year in at the time of the talk, and work mainly with insurance brokerages, helping them grow without growing their headcount by stripping out monotonous admin.

His path in was unusual. He moved to New Zealand in late 2019 and spent five years at Sacred Heart College in Auckland. He and Caique were both academic, on track for the typical law and finance double major, but in year 13 they became interested in sales and even planned to spend a couple of years as car salesmen to learn persuasion before starting a business. Mentors changed that: they met Andrew Patterson, and later Liam, who helped them along the way. Don started BlueHorizon early last year while also studying graphic design at AUT. He went in with zero business degrees and zero AI degrees. Caique learned an AI course with his dad, originally in Portuguese, and they taught themselves automation from there.

Their first projects were a free website chatbot for a physio clinic and a voice agent for lead qualification at a life insurance company. Both went all right, but the insurance broker told Don the real pain was internal admin, and that is where BlueHorizon focused.

The coolest internal use of AI is lead enrichment and personalisation at scale. For cold outreach they scrape a prospect's website, read how the brokerage positions itself, and rewrite the front-page copy into a personalised opening line that makes the prospect feel genuinely researched. The stack is an instant scraper to grab the data, into Firecrawl, then into Manus or Claude.

On the build side, the bread and butter is insurance onboarding, the form-heavy, back-and-forth process between client, broker and insurer, which they automate for both commercial and life and health brokers. They have also moved into voice, running a multichannel cross-sell agent across WhatsApp and voice: one batch of 108 prospects booked around 10 appointments in the first three days, and most people did not notice or mind that it was AI. Lead reactivation is another big one, re-engaging the thousands of cold contacts that businesses like real estate agencies let fall through the cracks.

His honest lesson is that content without substance does not work. Until the middle of last year he was posting YouTube videos and LinkedIn content about AI, but it had no real proof behind it, no specific this is what I did for a broker and here is the result. The advice he keeps coming back to is to do epic work first, then talk about it, because if you have not done the work you have nothing real to say.

What he would tell his past self is that one to three excellent client projects are your greatest asset early on. For BlueHorizon that meant picking up the phone. He shows a photo of himself and Caique in a Canterbury University study room with a cold call script in August last year. They did roughly five cold calls, five people picked up, three were interested, and they closed two. A few good months followed, enough to fund a trip to an AI event in Sydney.

His tips build on that: land one to three incredible client projects, even free ones, because a struggling business will usually say yes; do not keep that work in isolation, but tell as many similar businesses about it as you can, whether by cold call or large cold outreach campaigns; and dial in your sales process by tracking everything, so a no-show problem or an objection-handling problem becomes something you can actually fix with automation.

Key takeaways

• Pick one tool and go. Both speakers' biggest mistake was spreading across too many tools or projects instead of going deep on one.
• Set up automation properly and it is a no-brainer, even in creative fields like marketing. The key is solving a real, repetitive problem.
• Do the work before you talk about it. Content only converts when it is backed by real results for real clients.
• Early on, one to three excellent client projects matter more than anything, even free ones, because they build the proof and testimonials you need.
• Most leads come from conversations. Word of mouth, cold calls and outreach beat waiting to be found.