Should you hire or contract an AI specialist? Three routes compared

An employee of your own knows the organization from the inside, is there every day and builds up knowledge that stays. That's valuable, especially if AI becomes a permanent part of your work.

The downside:

• The people are scarce. Good AI specialists are in demand, so recruiting takes time and costs money.
• One person, one profile. A strong developer is rarely also a good trainer or adviser, and vice versa.
• Vulnerable. With illness, holidays or departure, everything stops, and the knowledge leaves with them.
• Keeping up is a job in itself. Models and tools change every few months. On your own, that's hard to keep up with alongside the regular work.

Fits when: you want to tackle lots of processes at once, you already have your own IT department, and AI becomes an ongoing stream of build work.

A freelancer is available quickly and often strong in one thing: an n8n specialist, a prompt engineer, a developer who builds a connection. For a well-defined job, that's exactly right.

Watch out for three things:

• Who manages it afterwards? An automation or agent needs maintenance: models change, and so do systems. Agree beforehand who does that.
• Where does the knowledge stay? Ask for documentation and a handover, otherwise the knowledge leaves at the end of the assignment.
• Who chooses the job? A freelancer builds what you ask for. Whether that's the process that delivers the most, you have to know yourself already.

Fits when: you know exactly what needs to be built, and your own IT can manage it afterwards.

A permanent partner provides a team instead of one person: someone who advises, developers who build, someone who trains and keeps an eye on the project. You buy capacity per month, and the partner makes sure the right people are at the table at the right time.

That's how our AI Partnership works: 30, 50 or 80 hours a month, for consulting and building at the same rate of € 150 per hour. It starts with the AI Quick Scan: one day at your site, and within a week a business case and a roadmap. We arrange hosting and maintenance separately, through a fixed maintenance contract.

The downside, honestly:

• It's a commitment. With us a minimum of 12 months, then 2 months' notice. That's deliberate: the first months are about the foundation, the return comes after that.
• The partner isn't in your building. You still need someone in-house who guards the priorities and brings colleagues along.
• For one job it's too much. If you only want one connection, a freelancer is cheaper.

Fits when: you want to tackle several processes, have no AI team of your own and want it to keep working after it's been built.

The role yes, the job title not necessarily. Someone has to own AI in the organization: setting priorities, guarding the AI policy, bringing colleagues along and directing the partner or specialist.

In most organizations that's not a new position, but an existing manager or team leader who frees up part of their time for it. We call that person the internal champion. They don't need to be technical. Curiosity, authority in the organization and time matter more.

You need the champion with every route. A specialist without an owner builds the wrong thing, and so does a partner without an owner.

Don't just compare the rate, but everything that comes with it.

• Hiring: salary plus employer costs, recruitment, onboarding time, training and tools. Also factor in that one person can't do everything, so sometimes you bring someone in on top anyway.
• Freelancer: an hourly rate, plus the time it takes you to choose the job, steer it and arrange the management afterwards. Rates vary widely; ask for several quotes.
• Permanent partner: a fixed amount per month. With us € 4,500 (S), € 7,500 (M) or € 12,000 (L), with the € 6,000 Quick Scan as the start. In the first year that's € 60,000, € 96,000 or € 150,000, excluding VAT. For what runs, a maintenance contract comes on top: € 245, € 445 or € 895 per month, including AI usage.

Which route is cheaper depends on what it delivers. That's why in the Quick Scan we first calculate what AI saves in your processes. Our guarantee: at least 2x back in the first year, otherwise we don't start. You can already run your own numbers with the AI savings calculator.

In practice, most organizations don't choose one route, but a sequence:

1. 01Appoint an internal champion. Someone who gets time for it.
2. 02Bring in knowledge with a partner. The partner advises, builds and trains, and brings the champion along in everything that happens.
3. 03Let the team grow itself. Colleagues learn to make the simple applications themselves: prompts, their own assistants. The build work stays with the partner.
4. 04Only hire someone later, if needed. After a year you know exactly which profile you're looking for, and that person has a foundation to build on.

That way you start quickly, without the knowledge depending on one person, and you keep your options open.