CUSTOM SOFTWARE

Software built around how you work

When the off-the-shelf tool almost fits, the gap gets paid for in manual work every single day. Custom software closes that gap instead of forcing the process to bend around a product.

What you get

Dashboards and analysis

Your numbers in one place, in the shape your business reads them – not the shape the tool vendor chose.

Connectors for AI models

One endpoint through which ChatGPT, Claude or any other model queries your numbers directly – read-only, aggregated, with a quota. Exactly that already runs in the Restaurant Dashboard.

Interfaces between systems

APIs and imports so data moves by itself instead of through copy and paste.

Internal tools

The screen your team needs daily, without the twenty features nobody uses.

An honest answer first

If an existing product solves it, we say so. Building something you do not need is bad business for both sides.

One example, in production

The Restaurant Dashboard is our own product: analytics across several locations, a side-by-side comparison and an AI analyst that turns the numbers into sentences. The screenshots show two demo locations.

Dashboard view “Restaurant analytics” with metric cards and a revenue curve

ANALYTICS

Key figures against the previous period, revenue over time, busy hours by weekday and hour.

Dashboard view “Compare restaurants” with a comparison table of two demo locations

LOCATION COMPARISON

Any metrics side by side, best and weakest location highlighted per row.

Dashboard view “Data analyst” with six analysis modules

AI ANALYST

Prepared analyses run by a language model on your own data – with the cost per run in plain sight.

What else is in there

Historical trends
Time-series analysis aggregated by weekday, week and month – patterns instead of comparing single days.
Product analysis
Best-selling products, product groups and toppings. Shows the slow movers as clearly as the bestsellers.
Customer analysis
New versus returning customers, cohorts, RFM segments and customer value – aggregated, with no personal data.
Explorer
Free-form ad-hoc analysis: combine measures and dimensions yourself, without SQL and without asking IT.
Alerting
Anomaly detection: every day per location against its own weekday baseline, largest deviations first.
Revenue forecast
Seven-day net revenue forecast from historical patterns and weather data.
Deliveries & portals
Delivery times from kitchen to handover, delivery portals against web shop and till, cancellation rates per channel.

How it works

  1. 01

    Understand

    We look at the actual process, including the spreadsheet everyone pretends is temporary.

  2. 02

    Smallest useful version

    The first release does the one thing that hurts most. It ships early and gets used.

  3. 03

    Grow with it

    Extensions come from real use, not from a feature list written before anyone touched it.

What it costs

Hourly from €80 with an agreed scope. Bigger projects get split into stages you can stop after – you are never buying the whole thing before you have seen the first part work.

Common questions

How long does it take? +

A useful first version usually takes weeks, not months, because it does one thing rather than everything.

What happens to the code? +

It is yours, in your repository, documented. You can hand it to any other developer.

Can you work with our existing system? +

If it has an interface or a database we can read, yes. If it has neither, we say that plainly before starting.