Help Desks Measure Fixing the Problem. He Built for Preventing It. (Jevon Le Roux)
Failed to add items
Add to basket failed.
Add to Wish List failed.
Remove from Wish List failed.
Follow podcast failed
Unfollow podcast failed
-
Narrated by:
-
By:
Twenty years ago the help desk was the big step forward: take the customer-service inbox and triage it into tickets. Jevon Le Roux's objection is that the whole category is built on the premise that something has already gone wrong — and that by then the customer has already noticed. Keeyu's bet is the opposite: detect the operational failure in real time, fix it automatically, and keep the ticket from ever existing. His forecast on the episode: in five years, "where is my order" will not be a question anyone asks.
The origin is specific. In 2020, at P.E Nation, a planned physical warehouse sale moved online under lockdown. Demand was intense enough that the storefront and the warehouse stopped syncing, so orders kept being accepted against stock that wasn't there. Nobody found out for a week or two — until customers started calling. His co-founder Tracy pulled the thread and found roughly a thousand orders with no stock behind them, then had to call and refund them one by one. Her ask afterward was the product spec: if I could have seen that in real time, I could have stopped it from existing.
They built it as an internal system, and only understood what they had when they ended up rebuilding it from scratch at a second company. When they took it to a well-known Australian ecommerce founder while raising an angel round, his verdict was that it can't be solved — too broken, too much variation between businesses. The three of them read that as the reason to do it rather than a reason not to.
Jevon is the non-technical one: an operator with two decades in retail and ecommerce, a third-time founder, alongside Tahir Rauf on engineering and Tracy Godtschalk on product. Keeyu went live commercially in July 2024 and raised a $2.3 million pre-seed led by rampersand in November 2025.
Host Jonathan W. Buckley of The Artesian Network and Jevon get into:
- Why he tells founders to fall in love with the problem and stay unattached to the solution — and how his own product definition moved from "prevent the ticket" to "automate the work"
- The warehouse-sale failure that produced the first internal early-warning system, and the second rebuild that turned it into a company
- Why an early-warning system alone made things worse: it surfaced issues customers had been churning over silently, and left teams to fix them by hand
- Why he claims there is no false alert: built on frontier models, run on a deterministic base model with if-this-then-that logic, which he puts at a 99.97% automation success rate
- What he says couldn't have been built two years ago, and why only the last six to twelve months made it possible
- Being the non-technical founder of an AI company: what he had to learn fast, and what he had to stop having opinions about
- Why he won't claim product-market fit at roughly 25 to 30 customers with almost no churn — the positioning inside the org, and who the user actually is, aren't settled
- What the real objection turned out to be: not trust or data access, but whether an ops team has the bandwidth for another project
- Founder-led sales all the way to about 33 closed deals, one SDR feeding the funnel, and the go-to-market coach who changed how he sells
- Splitting the company: engineering and product stay in Australia for the talent and the R&D support, go-to-market moves to the US — "built in Australia for the USA"
- Why proactive operations never penciled before: the unit economics of doing it with people instead of agents
Connect with Jevon
LinkedIn: linkedin.com/in/jevonleroux
Keeyu: keeyu.com
Follow the show
YouTube: youtube.com/@theartesiannetwork
Spotify: open.spotify.com
Apple Podcasts: podcasts.apple.com
Every episode, on every platform: WeBuiltItBecauseWeHadTo.com
More from The Artesian Network: artesiannetwork.com