Talentgrid
Recruitment cloud
Hire on evidence,
not on a hunch.
Talentgrid runs a deterministic rule-matching engine and an explainable AI ranking engine side by side. Every score is reproducible, every recommendation is auditable, and every company on the platform is fully isolated from the next.
50 / 40 / 10
Rule · AI · Recruiter weighting
0–100
Reproducible match score
9
Recruitment roles
Top matches
Senior HVAC Engineer
A. Rahman
Lead HVAC Engineer · 11y
S. Menon
Mechanical Engineer · 8y
L. Ferreira
HVAC Design Engineer · 7y
K. Osei
Building Services Eng. · 5y
Scores combine a deterministic rule result with a semantic AI result. Mandatory rules are always evaluated first and can never be overridden by the model.
Modules
Everything a recruitment team runs on
Each module is feature-flagged per tenant — a disabled module never appears in that company's navigation.
Requisitions & Jobs
Approval workflows from draft to approved, then jobs published internally or externally with versioned matching rules.
OpenCandidate Database
Parsed CVs, structured skills, experience and certifications, duplicate detection and a searchable talent pool.
OpenDeterministic Rule Matching
Hard rules that must pass, soft rules that add weight, and a reproducible 0–100 rule score with evidence per rule.
OpenExplainable AI Ranking
Semantic ranking runs alongside — never over — mandatory rules, with why-it-matches, gaps and CV evidence.
OpenPipeline & Interviews
Configurable stages on a Kanban board, structured interview feedback and an immutable stage history.
OpenOffers & Hiring
Draft → approval → sent → accepted, with a hire record, time-to-hire and recruitment cost created automatically.
OpenReports & Analytics
Funnel, time-to-fill, source performance, recruiter and agency conversion, AI usage and rejection analysis.
OpenRecruitment Copilot
Natural-language candidate search, generated job descriptions and interview questions — scoped to your tenant only.
OpenMatching engine
Two engines. One auditable score.
Candidates are never matched to a job automatically. The recruiter picks the scope — a date range, everything since a date, only new and updated candidates, a manual selection, or the whole database — so AI credits are never consumed by surprise.
- 1Tenant validation
- 2Candidate scope selection
- 3Basic filtering
- 4Hard rule evaluation
- 5Weighted rule scoring
- 6AI semantic ranking
- 7Combined score
- 8Explanation generation
- 9Result stored against the run
Weighted rules
Fairness by construction
Gender, religion, race, ethnicity, marital status, disability and age are stripped from the AI payload before ranking. Where information is genuinely missing, the platform states “Not found in available candidate information.” rather than assuming a failure.
Governance
Built for companies that get audited
Server-side tenant isolation on every query — IDs cannot be tampered with
Role-based access control across nine recruitment roles
AI never overrides mandatory rules and never invents qualifications
Protected characteristics are excluded from every ranking decision
Historical matching runs and rule versions are never overwritten
Immutable audit log with before / after values on every change
Recruiter workspace
Dashboard, requisitions, jobs, candidates, pipeline, interviews, offers and hires.
Open workspacePlatform console
Tenants, subscription plans, feature flags, AI configuration and the system audit trail.
Platform owner onlyFirst-time setup
Provision the Platform Owner, subscription plans and three fully populated demo companies.
Run setup