Enterprise use case · Workday + OpenAI + Microsoft 365
Candidate screening with human review
Recruiters repeatedly read unstructured applications to locate role-relevant evidence, while missing information is easily mistaken for a negative finding. Recruiters receive source-backed comparisons of applications against explicit role requirements, with decisions left to people.
Systems involved
What each system contributes
The model step runs against the AI provider you choose, including Anthropic, Google Gemini or a model you host yourself. Anything outside a pre-built action runs through an API, protocol or database call, or as Rules + Code in Python, Bash, Go, Java, .NET, PHP, PowerShell or JavaScript/TypeScript, with the same approvals and audit.
Architecture
One governed run
How it starts: A recruiter releases a Workday application batch and the approved role requirements for review.
Step by step
How the process runs
- 01
Define the review criteria
Read the Workday role and permitted application content, excluding fields the screening policy does not allow the model to use.
- 02
Extract relevant evidence
The selected model finds application passages relating to the stated requirements and marks missing or ambiguous evidence explicitly.
- 03
Build the comparison
The model proposes a review order using only the approved job-related criteria, with the evidence and gaps shown for each application.
- 04
Deliver the review pack
Microsoft 365 presents the comparison to the authorized recruiter without distributing applications to a general team audience.
- 05
Record the human decision
Only after recruiter approval does a separate Workday step apply the permitted application-status update.
Control
Human decision and exception path
Human decision
A named recruiter reviews the original application before deciding progression or rejection; model ordering never excludes an applicant from review. Changes to screening criteria require the designated recruiting owner's approval.
Exception path
- Missing or unreadable evidence returns to the recruiter as unknown, not a failed qualification.
- Sensitive attributes or inferred personal traits are excluded from the comparison and trigger review of the affected output.
- Unsupported rankings or citations are withheld until a recruiter resolves them against the source application.
Evidence
What gets audited
- Role requirements and screening-policy version.
- Permitted application references and cited job-related evidence.
- Model comparison and gaps visible to the recruiter.
- Human disposition and any approved Workday response.
Nexus runs on premises, in private or sovereign cloud, or air-gapped. Credentials stay in the vault, roles decide who can approve, and the execution record stays inside your deployment boundary.
Sovereignty and deploymentRelated
Patterns, connectors and next steps
- Workday-ServiceNow integration
Integration pattern · Employment changes become traceable IT work with named owners and completion evidence.
- Workday connector
66 pre-built actions
- OpenAI connector
25 pre-built actions
- Microsoft 365 connector
45 pre-built actions
- Human-in-the-loop
Where approval belongs in a running workflow.
- Order to cash across CRM, ERP and billing
Another enterprise use case.
FAQ
Questions about this use case
- Does Nexus replace any of these systems?
- No. Each system keeps its own records. Nexus runs the process between them and records each decision.
- How does the run start?
- A recruiter releases a Workday application batch and the approved role requirements for review.
- What if a step is not covered by a pre-built action?
- The model step runs against the AI provider you choose, including Anthropic, Google Gemini or a model you host yourself. Anything outside a pre-built action runs through an API, protocol or database call, or as Rules + Code in Python, Bash, Go, Java, .NET, PHP, PowerShell or JavaScript/TypeScript, with the same approvals and audit.
- Can the model reject candidates automatically?
- No; this use case supports recruiter review, and each progression or rejection remains a human decision.
- Does it infer suitability from age, nationality or other personal characteristics?
- No; the configured review is limited to explicit job-related evidence and excludes sensitive attributes and inferred traits.
Use case review
Discuss this use case
Bring your version of the process. We will map the systems, approvals and evidence it needs.