Skip to main contentSkip to footer
search
ENG
NL

We don’t only talk about AI.
We build with it.//

AI at itenium isn’t a marketing add-on. It’s how we build software, how we train each other, and how we think about the next decade of consultancy.
This page is a snapshot of what that looks like in 2026 — the tools we’re building, the talks we’re giving, and the community we’re growing around them.

What we’re [building]

We invest in the layer between AI models and production software: the tooling, conventions, and engineering discipline that turn a coding agent into a teammate you can ship with.

Itenium.{Forge}

A microservice chassis we’re building in the open, with AI assistance baked into the development workflow from day one. The point isn’t to demonstrate AI — it’s to get a production-grade chassis to a level of polish and consistency that would be uneconomical to reach without it.
Much of Itenium.Forge is public and under active development.


Skills Framework

Our internal harness for guiding AI coding agents through real engineering work — the structure that lets agents stay on the rails across long, multi-step changes instead of drifting into plausible-but-wrong code.
It’s what makes Forge buildable at the pace we’re building it. We refine the framework while we use it; it stays private by design.
GitHubarrow_forward

How we work with AI

We put the guardrails in before we let an agent near the codebase: the tests, linters, and review gates that catch it when it drifts. We run every tool on our own projects before recommending it to anyone. And we write up what didn’t work as readily as what did — the failures are where the useful lessons are.
It also changes how we decide. A proof-of-concept used to cost enough that teams would argue for weeks over which of three directions to take. Now we can vibe-code all three in a fraction of the time — throwaway code is fine when the goal is to learn — keep the one that actually holds up, and build that one properly. Analysis paralysis loses its excuse.

AI raises the bar for fundamentals

AI didn’t make the fundamentals optional — it made them decisive. When an agent writes the code, the scarce skill becomes knowing what good looks like and steering toward it. That’s why our training tracks matter more now, not less.

Architecture

AI is strong at a function or a class. At system shape, not so much — it won’t organically arrive at a sound architecture the way an experienced architect would. To move at AI speed you need humans who can set the right structure and make it stick — deterministically, with fitness functions like ArchUnit an agent can’t argue its way around the way it can reinterpret a written guideline. Our Architecture Track has run for four years; it’s paying dividends now.
Architecture Trackarrow_forward

Testing

Test suites used to erode under deadline pressure and uneven commitment. AI makes testing cheap — a single “Do TDD“ in an AGENTS.md gets you real coverage — and the ambitious setups (e2e, mutation, architecture, Pact tests) are suddenly within reach. What it can’t do is decide what’s worth testing, at what level. That judgement is still the human’s, and it’s exactly what our Testing Track teaches.
Testing Trackarrow_forward

AI Track

A year-long program of talks on AI, security, agents, and the engineering around them. It opened with a full-day, company-wide AI Bootcamp — six teams, developers and backoffice alike, shipping a real app in a day. Some sessions stay internal to the team; the ones we can open up, we do. For example AI & Security: The S in MCP Stands for Security was also presented at VISUG and devs.gent.
See the full trackarrow_forward

AI-tenium

A small, recurring knowledge-sharing community of IT professionals — mostly people outside itenium — who meet up and trade what they’ve learned about AI in practice. Backed by a private Slack workspace.
It’s invite-only and intentionally small. We’re considering opening it up — carefully, though: it works because everyone shares as much as they learn, and that’s worth protecting.

;AI Blog Posts

All AI blog postsarrow_outward