What we automate, concretely
We do not sell «digital transformation»: we build pieces of system that remove repetitive work and give back reliable data. The five areas we work on most often:
- Documents into data. Invoices, delivery notes, expense claims, orders by email: capture, extraction, arithmetic checks, posting into the management system, a linked archive.
- Integration between systems. E-commerce, ERP, warehouse, CRM, couriers: data moves by itself, with an overnight reconciliation and a readable list of rejects.
- Automatic checks. Comparisons between systems that should agree, with an alert when they do not. It is the cheapest intervention and often the most revealing.
- AI where it belongs. Extraction from messy documents, classification, draft replies, search across company documents — always with deterministic verification and human oversight above a threshold.
- Machine data. PLCs, sensors, meters: uptime, stoppages and causes on a single page, with the system still collecting when the network is down.
How we work
- Analysis (free). One call and a few precise questions: volume, time per unit, errors, exceptions. By the end you know what that process costs you today.
- A proposal with numbers. Scope, timing, price, running cost, the share of cases we expect to be automatic and how it will be measured afterwards.
- Built in pieces. The first link in production within weeks, running in parallel: the old method stays active until the numbers say it can be switched off.
- Delivery and measurement. One page with the automation rate, rejects and errors. Essential documentation, credentials in your name, accessible code.
When we tell you no
We say don't do it when the process is about to change, when volumes do not justify the spend, when master data needs cleaning first, or when the problem is organisational and software would only hide it. A project that does not pay back within 24 months has to be justified by something other than savings — capacity, quality, compliance — or it should not be done.
Who works on your project
MadTech is a compact engineering studio: an electronic engineer and a product manager, both CEOs, both on the project. You talk to the people building it, not to an intermediary. Our systems run every day in retail shops with fiscal printers, in agriculture for treatment traceability, in greenhouses with local control that works offline, and in a platform that calculates sales-network commissions.
Sectors we move well in
- Retail: till, stock, reordering, online sales kept in step.
- Small manufacturing and workshops: machine data, stoppages, traceability.
- Services and offices: document cycle, checks, reporting that updates itself.
- Agriculture: registers, dosages, stock, documentary obligations.