From job posting to compliant hire, in one platform
AI handles the repetitive work. Compliance runs automatically in the background. You make every decision that matters. The platform covers the full hiring lifecycle: sourcing, screening, contracting, payroll, and ongoing management.
35+
languages parsed
15K+
skill taxonomy
170+
jurisdictions
4
data regions
The workflow
Six stages, each automated where possible
The platform handles sourcing, screening, compliance, contracts, payroll, and lifecycle management. You keep the hiring decision and the day-to-day direction.
1. Define the role
Describe what you need in plain language β seniority, skills, time zone, engagement type. The AI agent drafts a complete job posting from that brief: structured requirements, screening questions, and a compliant pay band benchmarked against market data for the target country.
The posting is structured data, not free text. Skills are mapped to a taxonomy, seniority is normalised, and requirements are classified as mandatory or preferred. This is what makes the matching engine work β it compares structured profiles, not keyword overlap in paragraphs.
2. AI sources and screens
Candidates are matched against your role using anonymised feature vectors. Each match comes with a plain-language explanation: which skills aligned, where the experience fits, what the availability looks like. Flagged or uncertain cases route to a human reviewer β the system is architecturally incapable of auto-rejecting an application.
Matching runs without names, photos, or nationalities in the pipeline. This isn't an opt-in 'blind hiring' toggle β it's a structural property of the matching architecture. Bias reduction is built in, not bolted on. Research shows this approach increases diverse candidate advancement by 25-40%.
3. You interview and decide
Message shortlisted candidates directly inside the platform. Every candidate carries a reusable identity verification badge β 6,000+ document types from 200+ countries, verified in under 4 hours on average. Share assessments, collect team feedback, and schedule interviews in one place.
The verification badge is persistent. Once verified, a candidate doesn't re-verify for each new application. This reduces friction for repeat candidates and means employers always know who they're talking to before the first conversation. Verification uses biometric liveness detection to prevent identity fraud.
4. Compliant engagement, generated automatically
Based on the role, the worker's country, and whether you have a local entity, the compliance engine determines the correct legal structure: direct hire, employer of record, or contractor of record. Contracts are generated per jurisdiction β not translated from a global template.
German contracts include works council provisions. Brazilian contracts include CLT clauses with mandatory 13th salary and FGTS. Indian contracts account for state-level variations on top of central Labour Codes. French contracts include RTT provisions and are generated in French as required by law. Every template was reviewed by local counsel.
5. Weekly payroll, handled
Once engaged, payroll, tax withholding, social contributions, and statutory filings run automatically. Workers are paid every Friday regardless of which payment rail fits their country. Tax calculations handle everything from France's 45%+ employer-side social contributions to Singapore's CPF for citizens and PRs only.
The system supports bank transfer, Stripe Connect, Wise, and local mobile money rails (via Sokin) for regions where SWIFT and card networks don't reach well. Settlement is weekly, not monthly β predictable for the worker and simpler for accounting. Payment is guaranteed to the worker regardless of client invoice status.
6. Ongoing lifecycle management
From onboarding to offboarding, every stage is tracked: contract execution, payroll, time tracking, expense management, performance review scheduling, notice periods, and termination procedures. The compliance engine monitors regulatory changes in real time and flags any engagement that needs attention.
When a country changes its labour law β a new statutory leave entitlement, a revised social contribution rate, a new data protection requirement β the system picks it up and flags affected engagements. You don't find out about compliance gaps from an audit. Affected contracts are updated proactively.
AI accountability
How we keep AI accountable
AI accelerates the workflow. It never replaces your judgement on the decisions that matter. Every guardrail is structural β built into the architecture, not a policy document.
No auto-rejection
Every automated decision exposes a human-review action. The system surfaces recommendations and flags β your team makes the calls. The architecture cannot auto-reject an application. This is the EU AI Act's 'high-risk' standard applied as a default, not just where required.
No training on personal data
Candidate personal data is never used to train models without explicit, per-candidate consent. Matching models are trained on anonymised, aggregate patterns β not individual profiles. This applies to all data, not just data covered by GDPR.
In-region data storage
Identity documents and personal profiles stay in the candidate's home region's database (North America, Europe, Middle East & Africa, or Asia-Pacific). Your team sees what a candidate has shared for that specific application, not raw identity documents or data from other applications.
Full fee transparency
Every fee is shown before you commit: platform subscription, per-engagement costs, and any surcharges. No hidden markups on worker salaries, no surprise compliance charges, no per-country premiums buried in invoices. The total cost of employment is calculated upfront.
Explainable AI
Every AI output β match scores, classification decisions, contract terms β comes with a plain-language explanation of why that output was produced. You can trace every decision back to the inputs that drove it. This makes audits straightforward and regulation compliance demonstrable.
Full audit trail
Every decision, every document, every payment is logged and exportable. When regulators, auditors, or internal compliance teams ask, the answer is already in the system β not in someone's email thread. Exports are available in CSV, PDF, and structured JSON formats.
Technical depth
How the key systems work
For teams that want to understand the engineering behind the product β how matching, compliance, contracts, and classification actually work under the hood.
How the matching engine actually works
Traditional job matching relies on keyword overlap: if the CV contains the same words as the job posting, it's a match. This breaks down across languages, across seniority levels, and across different naming conventions for the same skills. Our matching engine works differently. CVs are parsed into structured profiles: skills are extracted and mapped to a taxonomy (React.js, React, ReactJS, and React.js 18 all map to the same node). Experience is measured in structured terms: years, seniority level, industry context. The matching is done on anonymised feature vectors β the system compares skill+experience signatures, not keyword strings.
How compliance encoding works
Each country's employment law is encoded as a jurisdiction pack: a structured representation of employment rules, tax rates, social contribution formulas, leave entitlements, notice periods, severance calculations, and classification tests. When you create an engagement in Germany, the system pulls the German jurisdiction pack and generates a contract with German-specific clauses: works council provisions (Betriebsverfassungsgesetz), social insurance (Sozialversicherung) at the correct rates, and termination protections under the KΓΌndigungsschutzgesetz. This isn't a template with variables β it's a generative system that produces jurisdiction-specific output from structured law.
How contract generation works
Contracts aren't translations of a single global template. The system starts from the target country's legal requirements and builds up: what must the contract contain to be valid in that jurisdiction? French law requires contracts to be in French. Brazilian CLT contracts must include specific CTPS registration details. UAE contracts must specify the limited/unlimited distinction. Japanese contracts must detail the scope of work (gyΕmu naiyΕ) explicitly. The generation engine assembles clauses from the jurisdiction pack, fills in the engagement-specific terms (compensation, role, schedule), and produces a contract that would satisfy local legal review.
How the classification engine prevents misclassification
The classification engine runs the engagement details through the target country's specific legal test. In the US, this is a multi-factor analysis based on IRS guidelines (behavioral control, financial control, type of relationship). In the UK, it's an IR35 assessment. In Brazil, it checks for habituality, subordination, personal performance, and dependency β the CLT's four pillars. If the engagement structure fails the local test (e.g., a full-time, exclusive, managed role being structured as 'contractor'), the engine flags it and recommends EOR instead. This prevents the most expensive compliance failure in international hiring: worker misclassification.
Under the hood
Built for compliance at the infrastructure level
Compliance isn't a feature added on top of an existing platform. It's in the database architecture, the matching pipeline, and the contract generation engine.
Four regional databases
Data residency is enforced at the infrastructure level, not the application level. North America, Europe, MEA, and Asia-Pacific each have their own Neon Postgres cluster. A candidate's personal data never leaves their home region. This exceeds GDPR's requirements and satisfies data localisation laws in jurisdictions like India and China that require or prefer local storage.
Jurisdiction packs
Employment law, tax rules, and compliance requirements are encoded as structured data β not free-text guidelines. Each jurisdiction pack is reviewed by local counsel and updated when the law changes. This is what powers automated contract generation and compliance checking. There are currently 170+ jurisdiction packs covering every major hiring market.
AI matching pipeline
CV parsing handles 35+ languages using a combination of LLM extraction and structured NER. Skills are extracted and mapped to a normalised taxonomy of 15,000+ skills. Experience is structured and compared against role requirements on equal terms, regardless of the original CV format. Matching runs on anonymised vectors β no PII in the pipeline.
Compliance engine
Worker classification follows local law: the IRS multi-factor test in the US, subordination tests in the EU, the ABN test in Australia, the CLT framework in Brazil, the IR35 assessment in the UK. The engine can't be overridden to classify someone as a contractor when local law says they're an employee. Every classification comes with an explanation and the specific legal test that was applied.
Questions, answered
Platform questions
Can I bring my own candidates to the platform?+
Yes. You can source through Corpshore's matching engine, bring candidates you've found elsewhere, or both. Either way, the compliance engine runs the same checks β classification, contract generation, tax withholding β regardless of how the candidate was sourced. There's no additional fee for bringing your own candidates.
How does the AI handle candidates whose CVs are in different languages?+
The CV parser handles 35+ languages natively. Skills are extracted and mapped to a normalised taxonomy regardless of the source language. This means a developer in Japan and a developer in Brazil are compared on equal structural terms β same skill taxonomy, same experience mapping β even though their CVs were written in Japanese and Portuguese. The parser handles mixed-language CVs too (e.g., technical terms in English within a Spanish-language CV).
What happens if a candidate disputes an AI decision?+
Every AI output comes with a plain-language explanation. If a candidate or an employer disputes a match score, classification, or flag, the explanation is the starting point for review. A human reviewer can override any automated output. The system logs the override and the reason, so there's an audit trail of both the original decision and the correction.
How quickly can I go from posting a role to making a hire?+
Average time from brief to published posting is under 10 minutes. Candidate matching begins as soon as the posting goes live. Identity verification averages under 4 hours. Contract generation is immediate once the engagement type is determined. End-to-end, the fastest hires happen within a week; complex multi-country engagements typically take 2-3 weeks.
Is there an API for integrating with our existing ATS?+
The platform exposes a REST API for role creation, candidate retrieval, status updates, and engagement lifecycle events. If you already have an ATS you'd rather keep using for some functions, the API lets you push and pull data without switching entirely. Webhook support is available for real-time event notifications. Documentation is available on request.
How does the platform handle the EU AI Act requirements?+
The EU AI Act classifies AI systems used in employment decisions as 'high-risk.' Corpshore's matching engine is designed to exceed these requirements: every automated output has a human-review action, every decision is explainable, the system cannot auto-reject candidates, and there is a full audit trail of every AI-influenced decision. Data protection is handled through regional storage, and bias reduction is structural (anonymised matching pipeline) rather than a policy overlay.
What happens during onboarding for EOR employees?+
The EOR onboarding process varies by country but typically includes: employment contract execution (in local language where required), social insurance registration, tax authority registration, bank account/payment rail setup, and any country-specific requirements (e.g., CTPS registration in Brazil, PF/ESI registration in India, works council notification in Germany). The platform tracks each step and notifies you of progress. Most onboardings complete within 5-10 business days.
How does the platform handle terminations across different jurisdictions?+
Termination procedures vary dramatically by country. The platform's compliance engine generates the correct termination process for each jurisdiction: notice periods (from 0 in most US at-will states to 7 months in Germany for long-tenured employees), severance calculations (from none in the UK to one month per year in Argentina), required documentation, and any regulatory filings. For countries where termination requires government approval (e.g., Netherlands, India for certain categories), the platform guides you through the process.
See it in action
Post your first role in under 10 minutes. Free on the Starter tier, no credit card required. The compliance engine, matching pipeline, and contract generation work from the first role.