Decision Agent: evidence-based decisions
An agent that reads the results from every stage, consolidates evidence, recommends decisions, and moves candidates through the process.
Executive summary
- Dedicated page for a product menu item focused on operation, governance and practical usage.
- Structured content to explain workflow, input data, outputs and supported decisions.
- Reserved spaces for platform-specific screenshots in real usage context.
Operational view
What the agent does
At the end of screening, the recruiter has assessment scores, interview transcripts, identity verification, and background checks per candidate, scattered across separate reports. The Decision Agent reads it all, consolidates it into a comparable view, and recommends who advances, who waits, and who exits, with the rationale.
The recommendation is never a black box: every suggestion points to the evidence behind it. The recruiter approves, adjusts, or reverses, and the agent executes the moves in the workflow, with the appropriate communication to each candidate.
Application
When to use
Comparing finalists
When several candidates have completed different stages and the company needs a comparable view to decide with rigor.
Decision volume
When dozens of processes run at once and advancement decisions pile up, slowing the entire funnel.
Decision audit
When the company must justify why each candidate advanced or not, with traceable evidence.
Step by step
How it works
Consolidates the evidence
The agent gathers assessment scores, interview reports, identity verification, and background check results per candidate.
Compares against the job criteria
Each candidate is read against the advancement criteria defined in the process design, weighted by competency.
Recommends with rationale
The agent suggests advance, hold, or close, pointing to exactly which evidence supports each recommendation.
Executes after approval
With the recruiter's sign-off, the agent moves candidates through the workflow and sends the appropriate communication for each case.
Operation
Data, outputs and workflow impact
Inputs
- Results from technical, cognitive, and behavioral assessments.
- Reports from the AI Interview and human interviews.
- Identity verification and background check status.
Outputs
- Side-by-side candidate view with consolidated evidence.
- Justified advancement recommendations per candidate.
- Workflow moves executed after human approval.
Governance
- No candidate is rejected without human approval.
- Every recommendation shows the evidence behind it.
- Complete decision trail for audit and compliance.
Platform
Suggested platform screenshots
Screenshot placeholder
Candidate comparison view
Insert screenshot of the side-by-side view with evidence per candidate.
Screenshot placeholder
Justified recommendation
Insert screenshot of an agent recommendation with the linked evidence.
Screenshot placeholder
Decision trail
Insert screenshot of a process's decision and move history.
FAQ
Frequently asked questions
Does the agent reject candidates on its own?
No. It recommends with rationale and only executes moves after the recruiter's approval. The final decision is always human.
How do recommendations avoid bias?
The reading starts from explicit criteria defined in the job design and objective evidence: scores, transcripts, and checks. Every recommendation is auditable and reversible.
Can we export the decision trail?
Yes. The history of recommendations, approvals, and moves is available for audit and can be exported.
Companies that trust Workdex
Decide with evidence, not a pile of tabs
See the Decision Agent consolidate a real process and recommend decisions with rationale.