
Assess a recruitment agency as an operating partner, not a source of profiles: verify its fees, screening, service levels, guarantee, reporting, legal responsibilities, and recent outcomes before comparing proposals. Quality-of-hire measurement exposes the gap between promises and proof: 89% of talent-acquisition professionals expect it to become increasingly important, but only 25% are highly confident they can measure it effectively.
That confidence gap comes from LinkedIn's Future of Recruiting research. A credible agency should answer your questions with recent data, documented workflows, and contract-level terms, not broad promises about access to talent. Ask first about fees, screening, and the replacement guarantee because these reveal whether the agency is aligned with a successful hire rather than maximum candidate volume.
For a CTO, VP Engineering, or HR Director at a Series A–C company, the first call should function as an agency audit. Whether you're hiring an individual specialist or building LATAM teams across engineering, product, data, and sales, insist on evidence that matches your role, seniority, and employment model.
Treat every material sales claim as something the agency must define, demonstrate, and document.
What do you need?
Choose the hiring path that fits
After reading "10 Questions to Ask a Recruitment Agency", most teams compare these options before deciding how to hire.
The most useful questions to ask a recruitment agency follow your buying decisions: commercial alignment, screening quality, delivery, operating ownership, risk, and proof. Broader resources such as MAJC's recruitment agency guide can support initial research, but your decision should rest on evidence specific to your search.
1. Question 1 What fee model do you use, and what will we pay at each stage of the hire?
Ask the agency to identify its pricing model before discussing candidate access. Get the fee basis, payment triggers, included work, and cancellation terms in writing.
For a contingency proposal, establish whether the percentage applies to base salary or total first-year compensation. For retained search, request the payment schedule and deliverables attached to each stage. For fixed-fee or Recruitment Process Outsourcing arrangements, ask exactly which roles, screening services, and search changes the quoted price covers.
A comparison of recruiting-agency arrangements reports contingency pricing of roughly 20%–25% of first-year pay and approximately 33% for retained search. Treat these as negotiation context, not a universal rate card or proof that a proposal is competitive.
Request a worked commercial example
Give the agency an actual open role and approved compensation band. For an account executive, ask it to show whether commission, signing payments, or equity affect the fee; for an engineer, ask what happens if you change the seniority requirement during sourcing.
Invoice triggers: Identify whether payment follows contract signature, shortlist delivery, offer acceptance, or the employee's start.
Scope changes: Document the cost of pausing, cancelling, reopening, or materially changing the search.
Successful-hire cost: Include assessments, background checks, replacement work, and other separately billed services.</li>
Use GENTY's fixed-fee recruitment guide to frame questions about scope and predictable pricing. Reject a proposal that doesn't let finance calculate the total exposure before signature.
2. Question 2 What is your replacement guarantee, and exactly what triggers it?
A replacement promise isn't the same as a refund. Ask the agency to walk through the remedy for a resignation, performance termination, failed start, and role elimination.
The contract should distinguish replacement, refund, and credit. It should also state when coverage begins, how quickly you must notify the agency, what exclusions apply, and whether payment disputes suspend protection.
Consider a hypothetical backend engineer who accepts an offer but withdraws before starting. Ask whether the agency restarts the search without another fee, whether the original invoice remains payable, and whether the replacement must meet the original assessment criteria.
Test the guarantee against difficult departures
Use GENTY's replacement-guarantee explanation as a starting point for contract questions, not a substitute for the signed agreement. Ask the agency to supply its actual clause and anonymized examples of how it resolved claims.
Qualifying events: Separate voluntary resignation, performance concerns, misconduct, restructuring, and changes to the job.
Replacement standard: Preserve the agreed seniority, skills, compensation range, and availability requirements.
Failure remedy: Specify what happens if the agency can't deliver an acceptable replacement.
Claim evidence: Request claim counts, outcomes, resolution times, and reasons for denied claims.</li>
Practical rule: Don't accept “we'll take care of it” where the contract should define your remedy.
Your decision threshold is straightforward: the guarantee must describe a usable process, not merely a period of coverage. If the remedy depends entirely on the agency's discretion, negotiate clearer terms before proceeding.
3. Question 3 How do you assess skills before a candidate reaches us?
Ask the recruiter to demonstrate the screening process using an anonymized candidate report. A list of technologies on a résumé doesn't establish whether someone can perform your job.

For a DevOps search, request evidence of how the agency evaluates incident response, infrastructure decisions, troubleshooting, and communication. For an SDR search, ask how it assesses prospect research, objection handling, and written outreach rather than relying on previous job titles.
Inspect the assessment, assessor, and decision
Request the assessment itself, scoring rubric, assessor's qualifications, and submission report. The report should explain what the candidate demonstrated, where evidence is missing, and why the agency recommends an interview.
Ask these follow-up questions:
Technical ownership: Who evaluates technical work, and what relevant experience qualifies that person?
Role alignment: How does the assessment map to your must-have competencies?
Communication evidence: How are written updates, live discussion, and distributed-team communication assessed?
Rejection reasons: Which weaknesses prevent submission, and how are borderline candidates handled?</li>
Review GENTY's skills-first hiring approach when defining the evidence you expect. Your engineers should calibrate the first submissions against the agreed rubric before the agency expands the pipeline.
Require demonstrated competence before accepting a candidate as qualified. A smaller shortlist with clear evidence is preferable to a larger list that transfers all screening work to your team. Don't set a universal shortlist-size target; set a role-specific standard for admission to your interview process.
4. Question 4 What are your median time-to-shortlist and time-to-fill, and how are they measured?
A promise of fast hiring is meaningless until the agency defines the clock. Ask for separate measures of time to technically validated shortlist, first interview, offer, and acceptance.
SHRM's recruiting benchmark reports a median time-to-fill of 39 calendar days for nonexecutive positions in its 2026 data, drawn from more than 4,600 organizations, and defines the measure from requisition to offer acceptance, according to its recruiting benchmarking research. Don't compare that benchmark with an agency's sourcing-only timeline.
Separate agency time from client time
Request recent medians for comparable roles, along with sample sizes and start and end definitions. Segment results by seniority, country, role family, and employment model rather than accepting a portfolio-wide average.
For a senior software engineering search, the agency should show how long sourcing and validation took separately from your interview scheduling, assessment review, and compensation approval. Use GENTY's time-to-fill metrics guide to clarify those distinctions.
Your service-level agreement should specify:
First delivery: When the first qualified shortlist is due and what makes it qualified.
Feedback ownership: Who responds to candidate reviews and interview results.
Reporting cadence: When the agency updates stage counts, blockers, and next actions.
Missed targets: When escalation begins and what corrective action follows.</li>

Accept a timeline only when its definitions match the comparison you're making. A faster shortlist isn't useful if your interviewers repeatedly reject candidates for competencies the agency agreed to validate.
5. Question 5 How do you handle candidate data privacy, consent, and background checks?
Before sharing candidate information, request the agency's privacy notice, data-processing terms, retention schedule, and technology-vendor list. Ask who can access résumés, assessment results, references, and compensation information.
Extend the question to AI-assisted sourcing, screening, transcription, and ranking. The IAPP's recruitment privacy and AI guidance provides a relevant basis for examining transparency, automated decisions, downstream vendors, and human accountability.
Follow a candidate record through the workflow
Ask the agency to demonstrate what happens from initial contact to deletion. Identify which systems receive data, which countries host it, how candidates are notified, and how the agency handles access, correction, or reconsideration requests.
Consent and checks: Require documented candidate authorization before background checks and ask counsel to confirm the applicable requirements.
AI use: Establish whether external tools retain candidate information or use it for training.
Human review: Ask who can override automated rankings and how that decision is recorded.
Auditability: Request decision logs, access controls, deletion procedures, and incident-response responsibilities.</li>
For a cross-border engineer, ask how the agency verifies employment history without collecting unnecessary personal information. GENTY's background-check guidance can frame those operational questions; the nonprofit background check company guide is an adjacent resource, not a technology-hiring compliance standard.
Don't accept “human in the loop” without meaningful override authority. Your threshold is a documented data workflow, accountable decision-maker, and legal review appropriate to the jurisdictions involved.
6. Question 6 Who will run the account day to day, and how will reporting work?
Meet the recruiter who will deliver the search, not only the partner selling it. Ask about relevant placements, current workload, sourcing support, and availability during your working hours.
Request a sample pipeline report and ask the recruiter to explain a stalled search. The useful evidence isn't a polished dashboard; it's whether the recruiter can identify the bottleneck, assign ownership, and describe the next corrective action.
Define ownership before kickoff
For a product manager search, your report should distinguish candidates being sourced, screened, submitted, interviewed, rejected, and awaiting feedback. It should also record why candidates leave the process and what the agency will change.
Ask the agency to document:
Named owner: Who manages candidate communication, screening, scheduling, and escalation.
Capacity: How competing searches affect the attention allocated to your role.
Working rhythm: A recurring pipeline review, feedback expectations, and decisions requiring your input.
Continuity: Who takes over if the recruiter leaves or becomes unavailable.
System access: Whether you can inspect a live pipeline or receive a structured export.</li>
A dashboard without owners and next actions is a status display, not an operating system for the search.
Set the reporting benchmark around decision usefulness. You should be able to tell what's blocked, who must act, and when the next update is due without reconstructing the search from email threads.
Reject an arrangement in which account ownership remains unnamed or substitutions happen without notice. Seniority in the sales meeting doesn't compensate for unclear delivery responsibility.
7. Question 7 How do you structure cross-border employment, and who carries the compliance risk?
Ask the agency to distinguish its recruitment service from the employment arrangement. Recruiting a person doesn't, by itself, establish who employs them, administers payroll, or bears classification risk.

For US engagements within its scope, the Department of Labor describes a six-factor economic-realities analysis and states that a contract's label alone doesn't determine worker status in its FLSA employment-relationship guidance. Don't treat that framework as a substitute for country-specific advice.
Require an employment responsibility map
Consider an engineer who works indefinitely under your product team's direction. Ask local counsel to assess the proposed contractor, direct-employment, or employer-of-record arrangement rather than accepting the agency's preferred label.
The written responsibility map should identify:
Contracting entity: Who signs with the worker and under which governing law.
Administration: Who handles payroll, withholding, benefits, leave, insurance, and required records.
Classification: Who assesses status and responds to a challenge.
Termination: Who approves dismissal and calculates applicable obligations.
Liability: What indemnities cover, what they exclude, and which limits apply.</li>
Request local legal guidance for every hiring country and compare total employment cost, not salary alone. Include administration, benefits, statutory obligations, currency handling, and termination exposure in the evaluation.
Your threshold is written confirmation from qualified advisers and clear contractual ownership. An agency's regional experience is useful, but it isn't a legal opinion or an unlimited transfer of your compliance obligations.
8. Question 8 Which salary data do you use for benchmarking, and how current is it?
Ask to inspect the salary benchmark behind the recommendation. You need its source, collection dates, sample size, role definitions, and treatment of rejected and accepted offers.
A generic software engineer category isn't enough for a specialized search. Define the job through scope, autonomy, technical depth, leadership responsibilities, and working arrangements, then ask the agency to explain how it mapped comparable candidates.
Compare equivalent compensation packages
For a platform engineer, distinguish base pay from bonuses, equity, benefits, and employer costs. For an account executive, separate guaranteed compensation from commission assumptions and ask how the benchmark treats attainment.
Request:
Data provenance: Whether figures come from placements, candidate expectations, surveys, or another dataset.
Recency: When the data was refreshed and what triggers an update.
Comparability: How country, seniority, employment model, and specialist skills affect the range.
Full cost: Whether employment administration and employer obligations are included.</li>
Ask European counsel to confirm current pay-disclosure and salary-history requirements, then require the agency to demonstrate how its workflow supports them. Don't accept a general statement that the process is compliant.
For a hypothetical search with repeated compensation-related withdrawals, the recruiter should distinguish an inaccurate range from changes in scope, candidate expectations, or your approval process. Require a defensible range and explicit limitations, not an unsupported claim that your budget is competitive.
If the underlying sample doesn't match your role, use the benchmark as directional input and validate it through the pilot search.
9. Question 9 How do you handle exclusivity, conflicts of interest, and candidate submissions?
Walk through a duplicate-submission scenario before agreeing to candidate ownership terms. If your internal recruiter already contacted a candidate and the agency later submits the same person, the contract should make the fee position clear.
Ask what establishes a valid introduction: a name, a résumé, candidate permission, or an agreed submission through your applicant tracking system. Also define how long ownership lasts and how prior contact is evidenced.
Make submission rules enforceable
Ask the agency to disclose sourcing restrictions arising from existing client relationships. Those restrictions matter if they prevent the recruiter from approaching employers that contain much of your target talent pool.
Candidate permission: Require confirmation that the person agreed to the specific submission.
Duplicate checks: Establish how the agency checks prior applications and introductions.
Exclusivity scope: Clarify whether exclusivity applies to the search, candidate introduction, or both.
Conflicts: Ask which existing relationships restrict sourcing or create competing interests.
Disputes: Define how records will resolve competing fee claims.</li>
For a senior data hire, ask whether the candidate is interviewing elsewhere and how the agency communicates competing deadlines without disclosing confidential client information. Don't confuse transparent availability reporting with control over a candidate's choices.
Accept exclusivity only when the agency provides a defined commitment in return, such as dedicated search capacity and measurable delivery obligations. A broad ownership clause without submission standards exposes you to disputes while giving you little operational benefit.
Your internal recruiting team and agency should share contact rules before sourcing begins.
10. Question 10 What outcome metrics can you share, and which recent clients can we contact?
Finish by asking the agency to prove that its screening produces hires who perform and stay. Request first-interview conversion, offer acceptance, retention, and hiring-manager satisfaction for comparable searches.
The SmartRecruiters Recruitment Benchmarks report reports that hiring teams conducted 42% more interviews per hire in 2024 than in 2021, while average time to hire rose 24%. Candidate volume and interview activity aren't sufficient evidence of effectiveness.
Demand definitions and inspect the cohort
Ask for denominators, measurement windows, response rates, and treatment of replacement hires. Request 90-day retention as a defined post-placement indicator, alongside role-specific productivity measures, consistent with the measurement framework discussed in SHRM's data-driven recruiting guidance.
For engineering, agree on onboarding completion, independent contribution, and manager-rated competencies. For sales, define ramp milestones and performance expectations before placement rather than judging success only by offer acceptance.
Use reference calls to test the agency's response when a search goes wrong:
Comparable experience: Speak with clients who hired similar roles under similar constraints.
Failure handling: Ask what failed, how the agency responded, and who owned remediation.
Metric verification: Confirm whether reporting matched the client's records.
Repeat engagement: Ask what the client changed before using the agency again.</li>
If the agency claims audited controls or certifications, request the relevant documents and scope. Reject cherry-picked testimonials as a substitute for cohort-level outcomes, and record missing evidence as missing rather than treating it as neutral.
10-Point Recruitment Agency Comparison Guide
Turn the Questions into a Buying Scorecard
Score every finalist from one to five on fee transparency, skills validation, comparable service levels, guarantee clarity, account ownership, reporting, compliance, candidate exclusivity, salary intelligence, and verified outcomes. Treat that scale as your buying method, not an industry benchmark.
Anchor each score to evidence. The lowest score should mean missing or contradictory documentation; the highest should require a clear written commitment supported by relevant examples or records.
Apply risk gates before comparing totals
Weight the categories according to your hiring plan. A specialist engineering search needs strong technical validation and compensation intelligence; ongoing hiring also requires recruiter capacity, reporting discipline, and a sustainable commercial model.
Don't let an attractive total score cancel a material failure. Reject any agency that refuses to document fees, guarantee remedies, data handling, employment responsibilities, or candidate ownership.
Use this buying sequence:
Collect evidence: Obtain contracts, sample assessments, candidate reports, pipeline reporting, and comparable outcome data.
Check references: Test how the agency handles rejected shortlists, withdrawals, missed deadlines, and guarantee claims.
Calibrate the pilot: Run a paid or otherwise structured search for a high-priority role under explicit scope and commercial terms.
Inspect delivery: Compare the first shortlist with the agreed competency rubric, communication requirements, and compensation range.
Make the decision: Set a decision date before accepting further presentations or expanding the engagement.</li>
Choose the operating model deliberately
For ongoing hiring, compare agency-led searches with an RPO model. Ask which arrangement gives you the right combination of dedicated capacity, delivery accountability, internal coordination, and total cost under a realistic hiring plan.
For an individual LATAM role, request current shortlists and written commercial terms through GENTY's hiring service. Evaluate those submissions against the same scorecard you use for every other finalist.
The pilot should resolve uncertainty, not create an indefinite trial. Agree on what success looks like, who approves it, and what happens if the evidence falls short.
Choose the agency whose operating commitments you can verify and manage. Access to candidates matters, but documented screening, accountable delivery, and enforceable commercial terms determine whether that access supports your hiring plan.
GENTY recruitment offers tech and sales recruitment, skills-first shortlists, RPO, and salary benchmarking. Visit GENTY recruitment with your role brief and ask for the screening evidence, delivery commitments, and written commercial terms you need to evaluate a search.
