My dad was a small business owner.
The business was moderately successful, but as a teenager all I saw were the long, long evenings — manual invoicing, calculating taxes, ordering supplies, the work that came after the work. It stuck with me; it wasn't a life I wanted. So I took the safe route — corporate, salaried, someone else's risk — and chose to build technology rather than manage people.
But in doing so, I think I was suppressing something the whole time: an entrepreneurial spirit, an energy to push boundaries, a desire to control my own destiny. dareena.ai is a manifestation of that need.
And there's the symmetry. We're building infrastructure that makes business easier for our partners and their users — that lets them focus on their business instead of getting lost in the ever-changing world of AI and voice agents. So I've come full circle: I turned away from a small business, became one, and in the process built the very tools I wish he'd had.
That's the whole idea, really. You don't configure complex tools — you configure your business.
We hide the hard part
It sounds obvious when you say it that way. It isn't.
Most voice AI platforms hand you a pile of primitives and wish you luck — orchestration layers, model endpoints, telephony, a speech stack — and leave you to wire them together and keep them from falling over. STT, TTS, VAD, turn detection, barge-in handling, P50 latency budgets, codec negotiation. The technology is extraordinary. The pace of change, relentless. But the experience of assembling it is exhausting for anyone not embedded in it. And it quietly excludes exactly the businesses that need it most: the ones without an engineering team to throw at the problem.
We took the opposite bet. Tenants write the personality and the strategy — what their agent should be, how it should handle a customer, where it should draw the line. The platform handles everything underneath: tool registration, the connections, the unglamorous plumbing. You describe the business you're running. We make it answer the phone.
The pivot that defined us
For a while we tried to be everything to everyone. Retail customers, wholesalers, developers — one platform, every door open. It's a tempting place to start, because saying no to a potential customer feels like leaving money on the table, especially when you're just starting up.
So earlier this year we made the call to focus: dareena.ai is wholesale-only. We serve the wholesalers, partners, and developers who build on top of us. They provision the businesses, own the customer relationship, put their own brand on the front. We became the layer underneath — deliberately, permanently.
It felt risky at the time. But narrowing clarified everything: who we build for, what we say no to, and where our strengths actually lie.
Making that call also forced us to confront something we'd been circling — what actually sets us apart was never going to be the voice stack. The models and the speech layers are commoditising fast, squeezed from one side by open-source infrastructure and from the other by the realtime APIs the big players are starting to ship natively. Anyone betting their moat on having the best STT-to-TTS pipeline is building on sand.
Our edge is elsewhere: in the productisation, in New Zealand data sovereignty and regulatory alignment, in the wholesale model, in the operational relationships we hold with the partners who trust us underneath their brand.
Why here
It would have been easier to build this somewhere else. Host cloud-based servers at minimal cost on the other side of the world. New Zealand is a long way from all of it.
But hosting data and services here was never a constraint I was working around — it's part of the point. When a business hands its customer conversations to an AI, it's handing over something intimate: who called, what they needed, what they said. Where that data lives, and whose laws govern it, isn't a footnote. For a growing number of our partners it's the whole decision.
So we made sovereignty a foundation rather than a feature. As of this week, our database runs on New Zealand soil. Sensitive data — recordings, customer conversations — stays here. Our compliance posture is built around New Zealand law, not retrofitted to it. When a partner asks where their customers' conversations are held, the answer is simple and it doesn't need an asterisk.
And the asterisk matters more than most people realise. Spinning up a major international provider's data centre that happens to sit in New Zealand isn't the same thing. If that provider is a US company, the CLOUD Act makes jurisdiction follow corporate ownership, not physical location. That's the difference between NZ data residency and NZ data sovereignty: one is about where the bytes sit, the other about whose courts can compel them. We chose the harder one.
Where we are today
Today the production platform goes live.
That sentence is short — but the effort, immense. The thing that lived in my head and on a hundred late-night branches is now real, running, answering calls for businesses that aren't mine.
Over the next weeks we'll be working with some great partners to put the platform fully through its paces. There's a particular feeling in shipping something you've carried this far — less fireworks than I expected, more a quiet exhale, ready for the next stage.
So if you're interested in joining us for this journey, reach out. We're excited to talk about how we can make it easier for you to grow your business in AI.
The last months have been some of the longest of my life — always doubts, always hurdles. But few experiences come close to the realisation that dareena.ai has a place in the maelstrom of AI, and that the value it brings to others is tangible.
It isn't one person, either. An idea can only be realised with others — so thank you, Kate, and Seeby, my co-founder — for your guidance, your experience, and your tolerance.
And Dad — thank you for being a role model. I know it's taken me a while to get here.