Cases: websites, online stores, bots and automation

What is here and what is not yet

This page holds the work we can show in full: websites for a dental clinic and a steel supplier, an online store for climate equipment and a Telegram bot for a meal delivery service. These are client projects carried through to a result, not portfolio concepts.

We work in more directions than the cases here suggest. Process automation and integrations live inside the business by nature — in the CRM, in messengers, in reporting, and in the work that no longer has to be done by hand. Their result is measured in freed-up time and enquiries that stopped getting lost, not in a hero image. That is harder to show in a screenshot than a website: the bot case carries a drawing of its own dialogue on the cover, because there is nothing to photograph. The rest of that work we still walk through on a call: what it was, what we automated, and what it produced.

What every case contains

The state before the work — not as decoration but as a baseline: what exactly got in the way, where money or time was lost. Then what we changed and why that particular solution, including the stack and the reasoning behind it.

And a measurable outcome. In the dental clinic case that is 70 to 90 clients from search in the first month with no advertising; in the climate equipment store, catalog load time down from ten seconds to 0.6; in the Telegram bot case, up to six hours of daily manual calculation the manager no longer does; for the steel supplier, five weeks from kickoff to launch. The numbers differ because the tasks differed: success is measured by the flow of enquiries, by speed and data integrity, by the time that was freed up, or by the delivery date.

How to map a case onto your own task

Matching by industry is the weakest signal of similarity. A dental clinic, a steel trader, an equipment store and a meal delivery service have nothing in common as niches, yet every task came down to the same thing: removing whatever stands between a person’s interest and the order — an anonymous doctor, a catalog with no way to narrow it down, a slow page, or a manual calculation capped by the manager’s hours.

What is worth comparing is different: whether the starting state looks familiar — a site on a website builder, a slow catalog, clients tracked in Excel, data scattered across systems — and whether you define the result the same way. If so, the case will tell you more about your project than a formally "same-industry" example.

What the cases deliberately do not promise

There is no promise to repeat the numbers. Those 70 to 90 enquiries in the first month happened in a specific niche, a specific city and against a specific set of competitors; carrying that figure over to another project would be dishonest. What does carry over is the approach: work out where the path to money breaks, and fix that rather than what is easiest to build.