Insights for management and IT leads
From software projects in mid-sized companies: integrations, modernisation, AI and what software in live operation really needs.
AI Readiness Check
Rate your company across six dimensions and instantly see your result, your bottleneck and next steps.
Ask Your Video
Nobody finishes the two-hour training, and the knowledge rots. How to turn every recording into a one-page summary and a searchable question-and-answer archive — locally, without a single byte leaving the building.
When your data is not allowed in the cloud
"Data protection rules out AI for us" is true for exactly one variant. Why local models, on-premise operation and RAG on your own server have long been enough for many SME cases.
Are You Talking to the Salesperson or the One Who Builds It?
Three people delivered the pitch, and none of them ever contributed a single line to the project. Why the split between sales and delivery decides the outcome — and which question makes it visible.
Good AI Fails on Bad Master Data
We used the best model available. It reliably found the wrong customer — because there were three of him, in three spellings, and none of them was wrong.
The Last Mile
Your prototype runs. Congratulations — you’ve just done the cheapest requirements analysis of your life. Now comes the expensive part: the production-ready fifth that costs four fifths of the effort.
Modernising Legacy Without a Big Bang
The boldest plan was to replace the 18-year-old ERP over a single weekend. It wasn't the smartest. Why legacy gives way in layers — and why the go-live date on which everything is new at once is the most expensive date in the whole project.
Build vs. Buy Is Not a Matter of Faith
"Building it yourself is always more expensive." True — until you add up the standard solution’s customisation effort over five years. Why the honest answer is usually a mix.
Bad requirements cost more than bad developers
The best development team I know once drove a project into the wall. It wasn't the code. It was one word that everyone understood — just each of them differently.
Excel Is Not the Problem
The one file nobody is allowed to touch, the one that half the company runs on, isn't a problem. It's the most honest map of your real process — if you read it before you replace it.
AI Starts with Permissions, Not with Prompts
The AI assistant answered every question reliably — including the one about colleagues' salaries. Why an AI project doesn't start with the model, but with the question of who's allowed to see what.
Integrations rarely fail because of APIs
A REST API was up in two days. The project still sat idle for three months — because nobody had decided who fixes the wrong record at night.
It's just a small change
It's not the change that gets underestimated, but everything hanging off it. Why an extra field in a form touches eight places in the system — and none of them found it small.
AI Won't Save Bad Processes
AI amplifies whatever is already there — in both directions. Why an automated chaos process isn't a better process, just a faster one.
Software Development with AI: How We Work with Claude Code
Claude Code writes code faster than a person can type. What ends up in your software is decided and checked by people. How that works, and what it means for timelines and fixed prices.
AI in Mid-Sized Companies: Use Cases, Costs and First Steps
A guide for managing directors and heads of IT: where AI adds value in mid-sized companies, where it fails, what it costs and what an orderly start looks like.
When Knowledge Lives in One Head: Securing Operations
When a key person is absent, it becomes clear how much operational knowledge lived only in their head and their Excel spreadsheets. How to recognise this risk and secure it with documentation and software.
Digitalising Processes in Mid-Sized Companies: A Guide in Six Steps
Digitalisation in mid-sized companies rarely fails because of technology. It fails because of unclear processes and projects that try to do everything at once. A guide from the first stocktake to day-to-day operation.
Automating Business Processes: Which Ones First?
Every company has a long list of processes it would like to automate. Three criteria show where to start, and what needs to be settled so that the automation holds up in day-to-day operation.
RPA or API Integration? What Pays Off in the Mid-Market
A software robot that operates input screens is quickly set up and trips over the next update. When RPA still makes sense, and why an interface is usually the better route.
Let's talk about your project.
In the intro call we clarify whether custom software pays off for you and roughly what it costs. Free and without obligation.
Book an intro call