Every Personalvermittlung quote starts with a clean-looking number: "20% of first-year salary" or "a flat fee of 18,000 euros." That number is real, but it is rarely the whole invoice. Between exclusivity clauses, replacement windows, notice-period technicalities and how the salary base itself gets calculated, the effective cost of a placement routinely runs higher, or structured very differently, than the headline percentage suggests. This is the full breakdown: the three fee models actually in use, what tech and AI roles cost in the German market specifically, and the charges that show up after you've already signed.
The three fee models a Personalvermittlung actually uses
Nearly every German recruitment agency prices its service one of three ways, sometimes blending elements of more than one. Understanding which model you're being quoted matters more than the headline number, because it determines who carries the risk if the search takes longer than expected.
- Contingency (Erfolgshonorar): the agency is paid only when a candidate is hired, typically a percentage of first-year gross salary. No placement, no fee. This is the most common model for mid-market tech roles, and it rewards the agency for submitting candidates fast, not necessarily for submitting the best-fitted ones.
- Retained search (Exklusivmandat): the client pays in stages, often a third at kickoff, a third at shortlist, a third at placement, in exchange for the agency's exclusive, prioritized effort on the search. This model is standard for senior, hard-to-fill or leadership roles where a contingency agency wouldn't invest real hours without a fee guarantee.
- Temp-to-perm (Try & Hire): the candidate starts as leased staff under Arbeitnehmerüberlassung, with an option (sometimes an obligation after a set period) to convert to a permanent hire, at which point a smaller conversion fee applies. This lets both sides test fit before a permanent commitment, at the cost of an ongoing hourly markup during the trial period.
What tech and AI roles cost specifically in the German market
Fee percentages vary by role scarcity, not just seniority. A rare AI/ML specialist commands a higher percentage than a similarly senior but more replaceable generalist engineer, because the agency's real effort (and the client's cost of getting it wrong) is higher.
| Role level | Typical fee (% of first-year gross salary) | Why |
|---|---|---|
| Junior/mid IT roles | 15-22% | Larger candidate pool, agencies compete on price |
| Senior software engineers | 20-28% | Standard range, negotiable above 25% with volume |
| AI/ML specialists, data engineers | 25-33% | Scarcity premium, genuine technical vetting effort |
| Engineering leadership (Head of AI, VP Eng) | 28-35%+ | Often quoted flat rather than as a percentage |
The costs that don't show up on the headline quote
- Exclusivity requirements: many agencies quote their best rate only if you commit to an exclusive mandate, running a parallel search with another agency or your own team can void that pricing entirely.
- Replacement guarantee fine print: a 3-6 month Garantiezeit sounds reassuring until you read what actually triggers it, voluntary resignation, performance-based termination and mutual separation are frequently treated differently, and some agencies only offer a free re-search, not a refund.
- Fee base calculation: whether the percentage applies to base salary only, or to base plus guaranteed bonus and sign-on, changes the effective cost meaningfully for senior roles where variable comp is a large share of total compensation.
- Invoice timing: some agencies bill on contract signature, others on the candidate's actual start date, if the candidate has a long notice period, that gap can matter for your budget planning and for what happens if they never actually start.
- Notice-period interaction: a candidate with a 3-month Kündigungsfrist at their current employer means a 3-month gap between "placed" and "working," which can affect both invoice timing and whether a replacement guarantee clock has already started running.
Comparing the three models side by side
| Contingency | Retained search | Temp-to-perm | |
|---|---|---|---|
| Upfront cost | None, pay on hire | Staged payments during the search | Ongoing hourly/monthly markup |
| Who carries search risk | The agency | Shared, client pays regardless of outcome timing | The client, via ongoing rate |
| Typical use case | Mid-market tech roles, high candidate volume | Senior, leadership or hard-to-fill roles | Testing fit before permanent commitment |
| Speed incentive | Fast submissions, not necessarily best fit | Depends on agency accountability, not fee structure | Not fee-driven, trial period sets the pace |
| Best for | Roles with an established candidate pool | Scarce, business-critical hires | Roles where fit risk outweighs urgency |
Where Aiporate's model differs, honestly
Aiporate is a new company without a long client history to quote statistics from, so this section is about structure, not track record. What we can say plainly: Aiporate's operating model is built around a 72-hour vetted-shortlist SLA and technical vetting done by people who have shipped real AI systems, not a generalist recruiter reading a CV. The fee structure is transparent and discussed upfront, without exclusivity fine print designed to lock you out of comparing options. If speed and genuine AI-specific vetting matter more to you than the lowest headline percentage, that is the actual trade-off to evaluate, not a claim about how many placements we've made historically.
Questions to ask before you sign, regardless of the model
- Is this fee based on base salary only, or base plus bonus and sign-on?
- What specifically triggers the replacement guarantee, and is it a refund, a free re-search, or both?
- Am I required to work exclusively with this agency, and does the quoted rate change if I'm not?
- When is the invoice due, on signature or on start date, and what happens if the candidate's notice period delays the start?
- What does the technical vetting process actually consist of, since a lower fee paired with weak vetting is the most expensive outcome, not the cheapest?
