Development
When the process will not fit
Three ways to get a system that fits your process: build it on a proven platform base, put AI into what already runs, or write it from scratch. Which one is cheaper in your case — we will tell you in the first conversation.
What we do
Three routes
They differ not in quality but in how much is already built before you start. Below — from the fastest to the most open-ended.
SaaS platforms
Your process starts on a base that is already running in production: CRM, HR records, recruiting, customer communication. Faster to start, more predictable to operate — and without the rigid limits of an off-the-shelf product.
CRM
HR
Recruiting
Communication
AI and LLM applications
A language model or machine learning built into an existing process: parsing documents and requests, assistance on recurring decisions, analysis of unstructured data. Part of the system, not a toy bolted on the side.
Automation
Assistants
Data analysis
EU AI Act
Custom development
Software written around the way you work, not the other way round. From the first requirement to production: AI speeds up the routine, architectural decisions stay with an experienced engineer.
MVP
Legacy
Integrations
EU hosting
Who it is for
When development
Not every need calls for custom software. If a ready-made product will do, we say so. Development pays off when at least one of these four is true.
Working around the tool
The off-the-shelf product makes you adapt to it rather than the reverse, and every workaround eventually becomes the rule.
Spreadsheets instead of a system
Spreadsheets, email threads and disconnected tools pile up because nothing quite fits.
Nothing on the market fits
One core requirement is covered by no vendor — and that is exactly what gives you the edge.
Growth is held back by software
The software holds growth back instead of carrying it.
And if your process does fit a ready-made product, we will say so in the first conversation. Then you do not need a project.
The open ledger
Off the shelf, platform
Four things on which these three routes differ most. Where the off-the-shelf product wins, it says so here.
Off-the-shelf software
Entry cost
A licence from day one, predictable per user — the cheapest start.
Time to live
Days — ready immediately.
Process fit
The process follows the product; gaps turn into workarounds.
Evolution and exit
The vendor sets the roadmap and the price; data export happens on their terms.
SaaS platform
Entry cost
Usually from a five-figure sum — well below a full custom build.
Time to live
Weeks, not months: the base already runs in production.
Process fit
Proven blocks plus your own part: adapted exactly where it counts.
Evolution and exit
It follows your needs, with fixed people rather than a ticket queue.
Custom build
Entry cost
An MVP typically starts in the five-figure range.
Time to live
MVP in 8–12 weeks.
Process fit
The system follows the process, including the inconvenient part.
Evolution and exit
Code and data are yours; knowledge transfer is part of the operate stage.
A read from project experience, with no invented precision. Timescales and figures depend on scope and integrations — the exact quote comes after we go through the requirement.
How we work
Advise, build,
The order is the same whichever of the three routes you take.
01
Advise
First we work out whether custom software is needed at all — with an honest comparison against ready-made options. If the project involves AI, we classify the risk class and obligations under the EU AI Act up front. Architecture, data model and integrations are planned only after that.
02
Development
MVP first, then extended iteratively — with milestones you can see and test. AI speeds up the routine, but model choice, prompts and fine-tuning stay documented rather than hidden in a black box.
03
Operate
After launch we stay involved: operations, monitoring, further development. For AI systems we track both answer quality and the cost of model usage. Knowledge transfer to your team is part of the work — so you are not dependent on us forever.
What you can rely on
Accountability,
Privacy by architecture
Data protection is designed in from the architecture up, not bolted on after launch. Personal data reaches a model only where it is lawful and genuinely necessary.
EU AI Act assessed early
Wherever AI is involved, the risk class and obligations are settled before the first line of code, not after launch.
EU hosting on request
Hosted on infrastructure inside the European Union when your data must not leave it.
Development in the EU, no offshore
A fixed team with named contacts, not rotating subcontractors and a ticket queue.
Frequently asked
Development —
What does development cost?
MVP — from a five-figure sum
How long does a project take?
MVP in 8–12 weeks
How long until an AI application is live?
Pilot in 8–14 weeks
Can an existing system be extended or replaced?
Yes — the normal case
Where is my data hosted?
EU — on request
Platform or custom build?
It depends
Let us size up your project
A no-obligation call: you describe the plan, we tell you honestly whether we can help and by which of the three routes.