Available for new engagements
Fractional COO and program lead who writes the AI agents himself. Not an advisor who leaves a report behind, but someone who takes the operation over, automates it, and hands it back.
In practice at
Sound familiar
Billing, onboarding, reconciliation, reporting: it works, but only because three people carry it in their heads. Every growth step costs another one. You know it could be different, but nobody has the time to rebuild it while it's running.
You've seen the demos. What you need is one process genuinely running in production, with ownership, error handling and someone who understands what happens when it goes wrong. Not a proof of concept that stalls after three months.
Your COO or program manager is leaving, or you never got round to hiring one. You need someone who can be useful tomorrow, without a quarter of onboarding.
Who I am
I'm Jarko Mesuere. I started fifteen years ago in support and systems administration, built a national tolling platform from scratch as infrastructure lead, then ran the engineering team and the roadmap of the largest parking app in the Benelux, and today I'm COO of a proptech scale-up. Through JarkoMes BV I do exactly that for other companies: getting an operation under control and taking the manual work out of it.
That order is the point. I didn't learn about technology by sitting in meetings about it. I ran Kubernetes clusters, wrote Terraform and rolled back releases at three in the morning, and only then learned how to steer a P&L, report to a board and build a team.
What I do today combines the two. I take operations over, I make visible what is actually happening, and I build the agents and automations that take the work off people's plates. The same person who chairs the meeting writes the code. That removes a translation step, and that translation step is exactly where most AI projects die.
How I can help
Each with a clear scope and a clear price. Most engagements start small and grow with whatever turns out to be needed.
A diagnosis of your operation. Where time, money and visibility leak, and what an agent could take over. You get a working cockpit on your real numbers plus a prioritised list of what it's worth.
One concrete process automated and in production. No demo, no proof of concept. Running, with monitoring, error handling and ownership that transfers to your team.
I take the operation or the program over. Steer it, build on it, and make sure it can run without me. Scope, duration and involvement are something we work out together, depending on what the situation needs.
How I work
Nobody buys the solution they think they need. I start by measuring and observing, not by proposing. The problem usually sits one layer below the brief.
Exactly enough governance to create clarity, not an ounce more. Process should enable speed, not replace it.
I don't automate processes on paper. I work inside the operation and build on it while it keeps going. What I deliver runs in production, not in a sandbox.
I leave an organisation stronger than I found it, not one that depends on me. Documentation, ownership and transfer are in scope, not extra.
Work
Cadenza / Acrely — proptech
My current role. As COO of Cadenza I own the full operation: commercial, onboarding, finance and engineering delivery. The company was growing faster than its own visibility into the numbers, and that's where I started.
My own product · vlotr.bevlotr.be
Not a client engagement, but my own product. It is here because it proves what I sell: I don't only advise on building, I build and ship. Vlotr keeps track of what a volunteer board decides, who does what, and what its activities actually earn.
Plugin2Drive / ParKKing
A public-private program for the Parking.Brussels residents permit system, spanning multiple organisations — from development teams to CEO and general manager level.
From practice
One day of building dashboards exposed four structural leaks in an operation nobody considered broken. What that says about measuring inside a scale-up.
Read on 29 July 2026 · 3 min readMost AI projects never reach production. Not because of the technology, but because of three organisational questions nobody answers up front.
Read on 15 July 2026 · 3 min readWhat I learned building an always-on agent that saw everything and meant nothing. On signal, thresholds, and why an agent has to remember what it already said.
Read onOnce a month: how to build and run an operation with AI agents. Specific, from practice, no filler.
One-click unsubscribe. I share your address with no one.
Thirty minutes, no sales pitch. I'll tell you honestly whether I'm the right person, and if not, who is.
Contact
Tell me what you are running into using the form, or email me directly. I usually reply the same day.
JarkoMes BV · BTW BE1029746555