The short answer
Procuring an AI recruiter does not start with the demo – it starts with the question of which task you want to scope. After that, you check three things: does your organisational context steer the conversation, can candidates know that AI is involved, and do you know how automated outcomes are investigated and corrected when something goes wrong. The final hiring decision rests with you as the employer.
Bring your vacancy and current recruitment process. In 30 minutes, we discuss the reach, qualification and division of responsibilities that fit your team.
Scope the task first, then test
A procurement demo shows you how the system works under favourable conditions. What you want to know next: does this also work for your role, your team, your way of selecting people?
Start with a fictitious vacancy and example answers – including unclear or contradictory responses – to see how the system handles them. A practical test with real candidates comes later, once the basic settings are right. Then compare lead time, workload and the quality of what you get back.
Fictitious example: three candidates each mention the required qualifications, but one gives an unclear answer to a key question. How does the system handle that? What is the outcome, and can you review and discuss it?
- What specific task do you want the AI to carry out?
- Can you start with a fictitious vacancy and example answers, including unclear cases?
- How will you compare lead time and quality against your current approach?
Organisational context steers the conversation
An AI conversation is only as good as the context that drives it. Does the system work solely with a standard job title, or can you bring in your own requirements? The language your teams use, the workload in a specific department, the competencies that cannot be taught – that is the context that makes the conversation relevant or keeps it vague.
At Red Rocket, the Cockpit platform shows you the candidate profile, the WhatsApp conversation, the reasoning drawn from their answers, and optionally a CV submitted by the candidate. You bring in the job requirements. You make the final hiring decision.
- How do you feed your specific job requirements into the system?
- Can you review which questions were asked and what the candidate answered?
- What does the system do with unclear or contradictory answers?
Who knows what, and who can step in?
Candidates deserve to know that an automated system is involved in their selection. Ask every supplier how and when candidates are informed, and what route exists if someone has a question or objects.
An automated initial assessment can influence whether a candidate receives an invitation. That makes two questions necessary: how can you review the outcome and its reasoning, and what is the agreed route when an answer is missing, unclear, or processed incorrectly? Agree this before you start, so you know who is responsible for investigating and correcting errors.
The final hiring decision rests with you as the employer.
- How does the system inform candidates that AI is involved?
- What route does a candidate have for questions or objections?
- How can you review and substantiate an automated outcome?
- Who is responsible for investigating and correcting errors?
Which data goes where?
Which personal data does the system process? Where is it stored, in the Netherlands, within the EU, or elsewhere? Which parties have access, and for how long is the data retained? Ask the supplier to demonstrate this, not just state it.
Use for model training is a point you have confirmed in the contract. Ask explicitly whether candidate data is used to improve or train models, and have this confirmed in the data processing agreement.
Also ask which ATS integrations and connections to your own advertising account are available, and under what conditions they are set up. For processing details and limitations at Red Rocket, we refer you to our privacy and AI page.
- Which personal data is processed, and on what legal basis?
- Where is data stored and who has access?
- What are the retention periods and how do you request deletion?
- Is data used for model training? Have this confirmed in the contract.
What EU regulation says about AI in recruitment
The European Commission regards recruitment applications of AI as potentially high-risk applications. Transparency about how a system works and human oversight of automated decisions are recurring points of attention within that framework. The Commission has published a regulatory framework that elaborates on this. Treat it as background reading for your procurement conversations, not as a compliance checklist to work through yourself.
How Red Rocket approaches this
Red Rocket works on a no cure no pay basis. For a successful placement, our fee starts at 11% of the gross annual salary; the exact percentage and associated conditions are agreed upfront. Media budget is agreed separately.
The initial qualification based on answers in the Cockpit platform is not a verification of qualifications and not automatic proof of suitability. It is a structured starting point for your own assessment, not a decision you simply adopt.
Would you like to see how this works in practice for your vacancy? We look at a fictitious candidate conversation and your job requirements together. No commitment, but a concrete picture of what you are buying. Book a meeting and we will get to work on your role.
Frequently asked questions
What is the difference between this article and the comparison of AI recruiter suppliers?
The supplier comparison helps you choose between platforms. This article takes a step back: what questions do you ask before selecting a supplier? It is a checklist for your own team, not an evaluation of individual tools.
What does working with Red Rocket cost?
Red Rocket works on a no cure no pay basis, starting at 11% of the gross annual salary for a successful placement. The exact percentage and conditions are agreed upfront. Media budget is agreed separately. The current pricing breakdown is on our pricing page.
How does Red Rocket explain AI involvement to candidates?
Candidates are informed that an automated conversation forms part of the selection procedure. There is a route for questions and objections. The specific processing details and limitations are on our privacy and AI page.
Can I as an HR manager intervene if a candidate conversation goes wrong?
You can review the answers and the automated outcome, and discuss any ambiguities. Agree before you start who within your organisation is responsible for investigating automated outcomes and correcting errors. The final hiring decision rests with you as the employer.
