Custom software, integrations between systems, internal tools and apps. And where it's needed, the software on the device itself too. With AI, that usually costs far less time and money than you'd expect: 2 days to work out what's worth building, 5 to build it. Bigger work is just as possible. Something is already running by the end, not a report in a drawer.
Jens Téblickfounder, Niklu
on site
€1,300
Two days to map your processes and systems and work out what’s worth building — using AI as a tool, not as the pitch. You get a ranked list, one small working piece on your own data, and the honest verdict if custom software isn’t the right step yet.
on site
€3,250
Five days on site: from analysis to working software — an internal tool, an integration or an app your team can use and maintain from day one.
Day rate
€650/day
For teams who want to keep building after the five days, or whose earlier software needs adjusting: individual days at a fixed day rate, with no new engagement to set up.
Introductory rate for the first 10 engagementsPrices excl. VAT.
Two days on site: watching over your shoulder to see what happens today, what data exists, and where time gets lost.
02A ranked list of what's worth building, using AI where it's the fastest route and plain software where that's enough. If the honest conclusion is that you shouldn't build yet, you hear that too.
03If it continues, the build phase follows: five days from analysis to working software, faster than usual because AI speeds up the build — with permissions and the data model deliberately designed from day one, not patched in afterwards.
04At departure your team gets the working knowledge, running on infrastructure you own — with individual days available for anyone who wants to keep building afterwards.
Background
Seventeen years of experience in IT: servers, networks, cloud, IoT devices in the field, and the applications that run on them.
It started in web hosting. That grew into a company keeping thousands of websites and applications online, on its own servers across multiple datacentres. That business was later folded into Lemon Companies, which he co-founded and which built web applications and custom software. In 2020 he handed the company over to new owners.
At the moment the centre of that work is Veton, a Belgian manufacturer of EV charging stations. We do the full technical build there: the mechanical design, the management software on the device, remote updates, the cloud platform, the app, the ERP integration, the dashboards and the reporting.
An engagement starts with two days of watching how the work runs today. If it is not a fit, you will hear so, and it costs you nothing.
Systems that are still running today.
fig. 01 — Purchasing and stock signals
Reorder points, safety stock and an ABC breakdown calculated from our own order history — no crystal ball, just a timely signal when something is about to run out.
Needed a foundation to build on: roughly a year of order history and known lead times.
fig. 02 — Devices in the field, from software update to dashboard
A large fleet of EV chargers in the field, running software on the device itself that we wrote and update remotely. A failed update rolls back on its own, so none of them ever gets stuck on someone's wall.
fig. 03 — Automatic lead and email follow-up
Follow-up flows built in n8n, running in a self-hosted container — editable later without commissioning new custom work.
The machine, the measurement setup or the device at your customer's site knows something. Your ERP, your planning or your accounting knows something else. In between, there's usually a person with a spreadsheet.
That gap stays because two kinds of people look at it. The software vendor knows your back office and stays away from the device. The installer knows the device and never touches your software. Both are right about their own half.
At Veton, we walk that whole chain ourselves, every day, from the enclosure through production into the ERP and afterwards out at the customer. For the electronics we work on existing platforms. Writing the software that runs on them, connecting them and keeping them updated remotely, we do ourselves. We don't design printed circuit boards from scratch. The same approach works just as well when there's no device involved at all.
More concretely, by situation: integrations between systems · for installation companies · the digital work order
At Veton, the work goes beyond the technology. We handle purchasing, plan production and track stock. Finance, reporting and customer follow-up are part of that too.
That difference shows in what gets built. A report that shows everything is easy to make and rarely useful. The question is what you do with it tomorrow: order or wait, adjust a price, call a customer. Software that doesn't answer that is just a pricier version of your spreadsheet.
Anyone who has worked with a stock list that was accurate and still said nothing builds the next one differently.
Most people who build fast with AI can't run the result in production. Most people who can run production don't build fast. Seventeen years of building companies and doing the engineering inside them. At Lemon, we kept thousands of websites and applications online, on our own racks, servers and network infrastructure, spread across multiple datacentres. At Veton, we still build and operate the software that runs on the device, the cloud platform, app and ERP integration. That's the other half of what Niklu delivers: permissions and a data model deliberately designed from day one, no silent failures, and software that keeps running after handover, on infrastructure that's yours.
Not every problem is a good candidate for custom software, or for AI as part of how it gets built. Too little history, a process that's about to change anyway, or simply no worthwhile result within budget: it happens. If that's the honest conclusion, you hear it in the first days, not after the engagement, and at no extra cost or separate advisory step.
Video call in your browser — nothing to install, no preparation needed, no obligations.