Small AI engineering studio in Antwerp. For Belgian operators putting AI into production. Workshops, audits, AI assistants, automation. Built with your team, in your stack, on your data. Owned by your team after we leave.
Most AI shops are old firms with an AI practice bolted on. We started the other way. AI-native by operation first, so what we build for you actually reaches production. Agents do the research, drafts, and synthesis. The founder does the judgment, the build, and the handoff.
We build, ship, and document. Plain code, real systems. The founder does the engineering directly. The studio is AI-native by operation: agents handle research, drafts, and synthesis, so the human time goes to the work that actually ships. Senior engineering at small-studio prices, because we run what we sell.
Before we build, we audit. What to skip. What to test. Where the data is thin. Where the workflow will break. Where the compliance question lives. The audit names what NOT to build, so the build budget goes to what works.
Scoped, documented, validated, handed off. Your team runs it after we leave. No consultant dependency. No black box. The AI works because someone on your team understands how it works.
Pick what fits. Workshops to get your team productive, or a build to put AI on a workflow that's stuck. Flat fees on first engagements, scoped before we start. Outcome-based or scope-defined retainers on continuing work. No time-based or activity-based retainers.
A two-hour hands-on session for teams that pay for ChatGPT but only opened the chat box. We build a working AI helper for one of your real workflows, in your stack, live by end of session.
One or two days at your team. You leave with a two-page memo: what to try, what to skip, what to build first. No retainer attached. Build the next step with us, with your team, or with someone else.
AI that answers questions and does work. Plugged into your docs, tickets, CRM. Replies with sources, qualifies leads, triages tickets, updates records. Whatever your workflow actually needs.
Workflows that quietly do the work between your tools. n8n pipelines that move data, draft emails, update your CRM, route documents. The boring repeat work, off your team's plate.
A discovery week, a build sprint, and a handoff. Each stage has a deliverable you can hold, and a number you can point at.
One week. We sit with the people doing the work, map the workflow, and pick what's worth building first.
Four to twelve weeks. We build with your team, not for them. Tested on real cases before launch, documented for handoff.
Documented, monitored, owned by your team. Then we expand to the next workflow, or step out cleanly.
One AI workflow live in production for a paying client since May 2026. Lead-gen assistant + content engine, fully automated, no human-in-the-loop. Measurements landing Q3. More cases as we ship them. No fake roster.
Toubkal Wanderers ran on inbox and word-of-mouth. M'hammed, the founder, was answering the same itinerary questions over and over, and publishing one SEO post a quarter, when he had a free Sunday. We built him a lead-gen assistant on the website and an n8n content engine that turns trip transcripts and trail logs into search-ready posts, end-to-end, no human-in-the-loop. Live since May 2026; results landing Q3.
Built our lead-gen assistant and content engine. Running since May 2026, fully automated, no human-in-the-loop on posts.
innops is led by Zakaria Basraoui. Two years shipping production AI in real businesses: computer vision, MLOps, the unglamorous parts. Now applying the same engineering rigor to Belgian operators moving from prototype to production. A small AI engineering studio in Antwerp, AI-native by operation: senior on every engagement, agents handle research, synthesis and drafts, vetted freelancers brought in when a project calls for them. No junior pyramid, no slide factory. Current capacity: one or two builds at a time.