Offers
There are two ways to work together: a small, fixed-price integration, or a full Scoped Pilot. Either way, the price is agreed before work starts.
Who this is for: companies roughly 20–200 people who are past the 'maybe we should look into AI' stage and want something real — not a 60-page deck, and not an unaccountable freelancer who disappears after the invoice.
Two ways to start
Starter Integration
From €1,500, fixed price
Connecting two or three systems — HubSpot, Notion, Zendesk, and similar — mapped, tested, and handed over with documentation your team can actually follow. Scope is agreed before work starts; so is the price.
Scoped Pilot
Priced to scope, fixed price
An assessment, a working pilot (an AI agent, an automation, or a trained team), and a follow-through period. Full detail below.
Build or buy?
For most non-core needs, an off-the-shelf subscription is enough — that's not a shortcut, it's ongoing maintenance and updates from someone else's team.
Even when the answer is "buy," my value is in picking the right tool and wiring it in so your team actually uses it — that's what the Starter Integration covers.
Building is worth it only where you need something differentiated — customization nobody else can copy. It never costs less than a subscription, but it buys an edge no one can take from you. That's what the Scoped Pilot is for.
What the Scoped Pilot covers
I work with people who already suspect that AI can change how their company runs, and who still have to decide where that change is worth making. Most of my time goes into data science, machine learning, and the algorithm-level research that sits underneath a real product decision. The rest goes into the slower work around those systems: strategy, coaching, and the product and process design that lets a small or mid-sized company actually use what gets built.
At SME scale I take on AI transformation as consulting and coaching, not as a platform rollout. That usually means sitting with the decisions that shape the next product or the next operating process, then helping the people who will live with those decisions get good enough at AI-first development to keep going when I leave. Some of that is one-to-one. Some of it is a small cohort — three builders or leaders at a time — so the conversations stay specific. I also build agentic systems when the job is to make a workflow run with less supervision, and I help companies stand up an AI team when the honest next step is hiring and structuring, rather than buying another tool.
The work I take is concentrated in a few domains I already know how to judge: media and advertising, finance and investment, and HR and recruitment. Those are places where I have enough context to argue about what matters, instead of learning the industry from scratch on the client's clock.
After the build
Every engagement ends with a clear handover — documentation, training, and a formal sign-off.
After that, two options: pay as you go for occasional fixes, or — if you expect ongoing changes — a small monthly retainer that covers a set number of support hours, billed separately beyond that. This is something we set up only if you want it, not a default add-on to every engagement.
What I will not do
I turn down enterprise-scale programs, data-platform builds, generative video work, and compliance or legal consultation. Those problems already have vendors and playbooks. I would rather spend the hours on calls that still need a person in the room.
If you are trying to decide whether any of this is the right kind of help, write with the decision you are actually stuck on — leave a short note .