Chapter one

It started with being the family IT department

It's a role a lot of people will recognise. You're the one who gets the call about the printer, the strange text message, the password that stopped working, the photos that won't send. Same questions, different week, often from the same person, because an explanation given over the phone on a Tuesday doesn't survive to Thursday.

So Matt started recording Loom videos. Answer it once, send the link, and the answer stays put.

That worked well enough that it stopped being a family thing. The videos went up on a YouTube channel, other people started watching, and requests came in for things he hadn't thought to cover. We tried PDFs too.

Chapter two

Every version hit the same ceiling

A video library can only answer the questions you've already thought of.

People's real questions aren't like that. They're specific and endless. Someone wants to know how to do a particular thing, on their phone, with their bank, on the screen in front of them right now. You can't record a video for that. And you certainly can't do it at 11pm, which is often when someone finally sits down and tries.

We kept coming back to the same question: how do you answer everyday tech questions simply, for anyone, at any hour?

AI was the first thing that could. The job became taking the newest technology available and making it feel like no rush, no jargon, no judgement.

That line ended up being the whole product. Oskoole launched, and it worked.

Chapter three

Then it happened twice more

We thought we were running one product. Two things changed that.

One of our CTOs built a research tool for our own use. It reported back weekly with anything relevant to Oskoole, grants included. It was so useful internally that it stopped being a tool and became a product. That's the Dissemination Digest, now in pilot at RMIT with researchers and students.

Then at the end of a startup pitch event, a funder who had worked in Indigenous communities came up to us and said we should adapt the assistant for First Nations communities. That became Ngarrgu.ai.

Neither was on any plan. Both worked. That's when we understood what we actually had: not one product, but a method that travelled.

eggn.io was formed to be the home for it. One team, one engine, and a different assistant for each community.

Why it's called eggn

The tidy answer is eggs. Ideas, incubation, things hatching.

The real answer is better.

Matt took a month off to build the first MVP version of Oskoole. He'd coded CNC machining centres back in his twenties, so he wasn't starting from nothing, but this was another language entirely. Desk at 7am, coding till midday. Break. Back at it till dinner.

At the same time he was looking after a few neighbours' places. Hobby farms, owners mostly up in Melbourne. Cows to check, a pony to feed, and every one of them had chooks.

So come lunchtime he'd push back from the desk and say he was off to check the chooks. Which got shortened, the way these things do, to "I'm going eggn."

Someone asked where he was off to one lunchtime. That was the answer. The name stuck before anyone had decided it was a name.

We kept it because it's honest. A company that talks about no jargon shouldn't have a name assembled out of jargon. And nobody has ever remembered a name that came out of a branding workshop.

What we believe

Digital exclusion is a design problem, not a skills problem

People aren't failing to learn. They're being handed tools built for someone else. Exclusion is something we built, which means it's something we can un-build.

Nobody wants to ask twice

The barrier isn't understanding. It's the fear of looking foolish, and the sense that you're taking up someone's time. People stop asking long before they stop wondering.

Accessibility is the constraint, not the checkbox

One step per screen. Voice on by default rather than buried in a settings menu. Steps you can stop and start, download for print, and share with others.

Build with communities, not for them

Ngarrgu exists because someone who had worked in those communities told us it should. We didn't decide it from an office.

The team

The team is the reason this works, and it's the part we're proudest of.

Between us there are masters degrees in branding, gaming and interconnectivity, a PhD in progress in mental health literacy, an AI solution engineer based in Colombia, and a masters in data and computer science.

That's deliberate. When you're building for people who have been designed out of technology, you want game designers, brand people and mental health researchers in the room, not only engineers.

Matt and Martine Leete co-founded it. Between them they bring more than 30 years across manufacturing, retail, wholesale, ecommerce and international sales through the UK, France and the US.

What success actually looks like to us

Commercial success matters. It's what makes this sustainable rather than a grant-funded project, and we're not shy about that. But two things matter more.

After all these years in business, the thing that made it worthwhile was never the money. It was making a difference.

Decades of manufacturing, retail and international sales taught us plenty about building a company. What we took away from it is that the deals are not what you remember.

We're towards the end of our careers. Our team is at the start of theirs.

Getting to build something alongside people who are just getting going is humbling, and it is what makes this particular journey special.

Where this goes

A portfolio of AI assistants, one for each community mainstream technology keeps missing, all sitting on the same engine so that serving the next community costs a fraction of serving the first.

We want to be the obvious answer when an organisation asks how it can actually serve a group digitally. Not a consultancy. A working platform you can have branded and running in a week.

And we want to prove that AI built for social good can be a real business. If it only works with funding behind it, it doesn't scale, and if it doesn't scale it doesn't reach the people who need it.

Someone watching us present summed it up better than we do:

You're using AI for good, not evil.

That's the job.

Want to be part of it?

We're looking for pilot partners in aged care and retirement living for Oskoole, and First Nations organisations for Ngarrgu. If you serve a community that mainstream technology keeps missing, we'd like to talk.

Start a pilot conversation See Oskoole in action

Our story