Rebuilding globalsoftware.ai around the work

Most company websites are built once, polished for launch day, and then quietly drift away from reality. Ours was no exception. The site said one thing; the day-to-day work said another. So we rebuilt it — and this post is both an announcement and a table of contents for what you'll find here now.

What changed

The new site is organised around four things, because that's how the company actually operates:

Services. Global Software AI started as a consultancy and that remains the core: AI integration and training, security consulting, and custom software development. The services pages now describe the actual engagements we take on — from AI strategy for leadership teams to hands-on implementation — rather than a generic capabilities list.

Products. We build software of our own, and we'd rather show it in progress than pretend everything ships finished:

• CompylotAI is our compliance platform for financial institutions operating under EU regulation. It is in early access, and we're building it with the people who will use it. If your team reviews marketing materials, client communications, or internal policies against regulatory requirements, we'd like to talk.
• NectarForge is the odd one out: a mead-brewing app, launching soon. It exists because we believe small, real products are the best way to keep our development practice sharp — and because one of us brews mead.
• AI Employees, our managed AI assistant product, continues under its own roof at /agent.

Labs. Experiments, prototypes, and things that may never become products. We'd rather publish them than let them rot in a private repo.

This blog. Which brings us to the point.

Why a blog, and why now

Two honest reasons.

First, we keep having the same conversations with clients — about where AI genuinely helps, where it's oversold, what EU regulation actually demands of software, and how a small team ships production systems with AI assistance. Writing those answers down once, properly, is more useful than repeating them in calls.

Second, the way people find expertise is changing. Increasingly, answers arrive through AI assistants that cite their sources. We want to be a source worth citing — which means publishing clear, specific, honest writing about the things we genuinely know: compliance software for regulated financial services, practical AI adoption, and security. No growth-hacked listicles. If a post exists, it's because we had something to say.

What to expect

Posts will cover:

• Building CompylotAI in the open — the design decisions, the regulatory constraints, and what "human-in-the-loop" means when the stakes are real
• AI-assisted development as we actually practise it, including where it fails
• Security and compliance topics for teams operating under EU rules
• Occasional notes from Labs and from NectarForge's launch

We'll also publish video posts — recorded walkthroughs and talks, embedded here with written summaries so they're useful whether you watch or read. Those are coming; the writing starts today.

A note on language

The site is available in English, German, Latvian, and Russian, but blog posts are published in English only. Writing well in one language beats writing adequately in four, and our technical audience overwhelmingly reads English. The rest of the site remains fully translated.

Where we're writing from

Global Software AI is based in Riga, Latvia, founded by Tim Crouch after three decades across IT, cybersecurity, and financial services — including work with major banks and an EdTech company that grew to a nine-figure exit. That background shapes everything here: we build for regulated, high-stakes environments, and we write about what that actually takes.

If any of this is relevant to what you're building, get in touch. And if you'd rather just read along — welcome. The feed is at /rss.xml.