Human oversight, every tier
No client- or candidate-facing decision goes out without a named human’s approval — including in our highest-automation commercial tier.
ExploreTrust & governanceBuilt for sectors
Built for sectors
where trust isn’t optional.
These are the commitments that hold across every SOS Global system, regardless of division or commercial tier. We’d rather state them plainly here than bury them in a footer link nobody reads.
Public commitments
Six commitments, held across every system.
Consented data only
Where our systems draw on personal or candidate data, it comes only from data you already hold and have lawful consent to use. No open-web scraping, ever.
ExploreCompliance verified live
Where a placement or decision depends on a compliance status, it’s checked against the current, authoritative source at the point it matters.
ExploreIndependent challenge, structurally
Every system includes a function whose job is to find the reason an output might be wrong — built in, not bolted on.
ExploreIsolated by design
Where multiple clients or licensees use the same platform, their data is structurally isolated from one another. No bridging, ever.
Safe by default
New capabilities and integrations run in a fully simulated mode until proven safe — nothing goes live against real data before it’s ready.
On automated decisions
Where AI assists, and where a human decides.
UK data protection law gives individuals specific rights where a decision is made about them by automated means alone. Our position across every division is the same regardless of what the law strictly requires: a named human reviews the case, has genuine authority to change the outcome, and is competent to do so — not a name attached after the fact to a decision the system already made. Where we say “human sign-off,” that’s what we mean by it.
We publish this stance because it’s the single most common question serious buyers ask before an NDA conversation even starts. If you need more — the specific review points, the escalation paths, the audit trail — that’s exactly the kind of technical detail we share directly under NDA.
Due diligence register
Evidence buyers should be able to request.
Data map and roles
Approved purposes, categories, sources, processors, access and retention.
Security model
Environment boundaries, access control, logging, incident handling and testing.
Decision record
Inputs reviewed, exceptions, human action and accountable outcome.
Change control
Acceptance criteria, release approval, rollback and monitoring.
Claims evidence
Owner, source, date, methodology and expiry for every material public claim.
Buyer questions
What serious buyers ask before an NDA.
Is AI recruitment safe under UK GDPR?
It can be, when human review is genuinely meaningful — not a rubber stamp on an automated outcome — and when candidates are told plainly what role AI plays. SOS Global Recruitment requires a named human’s sign-off before any output reaches a client, in every tier.
Does SOS Global AI scrape candidate or personal data from the open web?
No. All candidate sourcing works exclusively from a client’s own consented CRM — never the open web, and never a purchased list.
How is compliance kept current rather than just checked once?
Through a dual-checkpoint model: every required item is checked once at screening, then re-verified against the current, authoritative source again on placement day. See recruitment compliance ↗ for the full detail.
Your next move
Clarity earns the next conversation.
These are commitments we hold ourselves to, stated plainly. If you’d like the underlying detail of how any of them is enforced — for due diligence, procurement, or framework audit purposes — we’re glad to share it directly under NDA.
Related AI governance intelligence
Continue your research.
AI governance and accountable decisions
Understand the controls behind evidence-led, human-accountable systems.
The SOS Global AI standard
See the principles applied across recruitment and entertainment work.
EU AI Act business compliance
Prepare operational evidence for changing AI governance obligations.
AI systems and authority resources
Move from policy language to inspectable controls.
Business AI systems
Configure evidence-led automation, bounded agentic workflows, exceptions and human authority around a defined business process.
Explore AI systems and business automation ↗Candidate identity controls
Connect application, interview, verification and onboarding identity with proportionate fraud controls and review.
Use the candidate identity checklist ↗Agent governance
Apply identity, permissions, monitoring, intervention, revocation and audit to AI agents.
Use the governance framework ↗