Skip to content
Focus area

AI-assisted software development for mid-sized companies

AI tools have brought custom software back within reach of mid-sized budgets. We explain what that means for your project and where the limits are.

The renaissance of custom software

For years, tailor-made software was too expensive for many mid-sized companies. They bought standard software and bent their own processes to fit it. AI tools have changed that calculation: routine work, which used to make up a large share of the effort, has become dramatically cheaper. What remains, and becomes more important, is what requires experience: understanding requirements, designing data models, security, operations.

What AI concretely does for us

Writing code that we review. Our developers work with Claude Code as a pair programmer. The tool produces drafts, routines, interfaces and migrations. The developer reads, corrects and decides. This website was built that way.

Tests that nobody else writes. Test cases are among the first things dropped under time pressure. AI-generated tests, reviewed by us, provide coverage that used to be unaffordable. Bugs surface before they reach the customer.

Reviews and security. In addition to human review, every change is checked automatically: for logic errors, for known vulnerabilities, for deviations from our conventions.

Documentation. Decisions, interfaces and operating instructions are maintained alongside the code instead of being written up at the end of the project. For you as the client that is the piece of independence many agencies never deliver.

Analysing legacy systems. In Oracle Forms modernisation and migration to PostgreSQL, AI helps read and translate code that is decades old. The decision about what to keep is ours.

What AI does not do for us

It does not design architecture, it does not decide about your data, and it does not replace a conversation with your business department. We promise no multiples and no miracles. We promise that you know where AI was involved and that an experienced human stands behind every result.

AI inside your application

Beyond the development process, we build AI features into applications where they have a clear benefit: recognising and filing documents, extracting data from invoices and contracts, structuring free text, assisting users with input. EasyFlow Business classifies business documents automatically this way. What matters to us is that the feature remains traceable and that your data stays on servers in Europe.

Frequently asked questions

Will my software be written by an AI?

No. It is built by our developers, who use AI tools such as Claude Code for code generation, tests, reviews and documentation. Architecture, data model, security and every decision about your requirements rest with people who have more than 25 years of experience. Every line is reviewed before it enters your system.

How much faster or cheaper does my project get?

Considerably, but we deliberately do not quote multipliers. The gain depends on how well the requirements are clarified and how much routine code the project contains. What we can promise: you see progress continuously in a test environment and are billed by actual hours, with time tracking you can inspect.

What happens to my data and my code when AI is used?

We work with providers whose contracts exclude the use of data for training, and before a project starts we agree with you which parts of the code and which data may go to AI services at all. Your customers' production data does not belong in an AI prompt. We regulate that contractually and technically.

Can you build AI features into our application?

Yes. Classifying documents, extracting data from invoices or contracts, summarising text, assisting users with input. In EasyFlow Business the application classifies 14 document types automatically via the Claude API. We make sure such features remain traceable and that a human has the final say where it matters.

Is AI-generated code maintainable?

Only if someone with experience fits it into a clean architecture and reviews it. That is exactly our job. Unguided, AI quickly produces lots of repetitive code that is hard to change. Guided, it produces code with tests and documentation that is more maintainable than much of what used to be written by hand.

More focus areas

Oracle APEX

New development, extension and support of APEX applications. PL/SQL in the back end, in use since 2005.

Oracle Forms modernisation

Inventory, upgrade or migration to APEX and open source. We have known Forms since 1998 and made the transition ourselves.

PostgreSQL & Oracle migration

Schema and PL/SQL migration to PostgreSQL, Go back ends, PostGIS. We know both databases and tell you when the switch pays off.

Hosting & operations

Operation of your application on European servers: monitoring, backups, updates, a named contact. Without dependency on cloud giants.

Let's talk

Tell us about your project.

Discuss your project