Pack 03 / Human Resources

HR with AI,from req to ramp.

JD generation, candidate screening, onboarding journeys, and a policy-aware helpdesk that deflects two-thirds of T1 tickets. Inside your HRIS, calibrated to your job ladders, governed by your policy.

Book a call
first deploy≤ 30 days
deflection68% T1
complianceLFPDPPP · SOC 2
maintenancemonthly evals
stackyour HRIS · your LLM
01Process Map

Every step, from intake to outcome.

We map your function, then deploy where the leverage is. The map below is the production pattern, not a workshop sketch.

01Recruiting02Onboarding03Performance04L&D05Helpdesk
02What We Deploy

Skills, agents, templates.

Three deliverable kinds, each scoped to your data and your tools. Setup is a one-time build. Maintenance is monthly, by us.

Skills
JD Generator · Resume Screener · Interview-Kit Builder · Policy Copilot
Setup + Maintenance
Agents
Candidate Screening · Onboarding Assistant · HR Helpdesk
Setup + Maintenance
Templates
Role-calibrated rubrics · Policy-aware Q&A guardrails
Setup
03Live Walkthrough
HR HELPDESK / 04.28LIVE QUEUE
T1 INBOUND
24/ hr
Deflected16
Routed to human8
Deflection rate67%
INTENT BREAKDOWN
  • Vacation accrual38%
  • Policy lookup28%
  • Onboarding step14%
  • Comp · routed12%
  • Conflict · routed8%
SCREENING RUBRIC / SR-MGR · Operationsmin 3.5
Communication
4.0
Domain knowledge
4.5
Leadership
2.8
below threshold
Cultural fit
4.2

HR Helpdesk Agent, T1 deflection.

Six hundred employees, twelve policy domains, one helpdesk. The agent answers in your tone, cites the policy section, and escalates anything ambiguous to a human within the same Slack thread.

  1. 01Employee asks about vacation accrual mid-cycle. The agent reads the policy, computes the precise figure from HRIS, and answers in two messages.
  2. 02Anything outside scope (compensation, conflict, disability) is flagged with a guardrail and routed to a human within the same thread.
  3. 03Every interaction is logged with the policy citation, the LLM call, and the resolution path. Compliance can audit in one query.
04Onboarding Journey

From kickoff to monthly maintenance.

Six weeks to production, then continuous coverage. The same motion every time, run inside your stack.

Week 1–2

Diagnostic + policy audit

Week 3–4

Skill/agent build + HRIS integration

Week 5–6

Pilot with live tickets + evals

Ongoing

Maintenance · policy updates

05Outcomes

What you can expect to measure.

Quantified outcomes from comparable deployments. Anonymized; conservative; tied to a specific 30-day window.

First deploy
≤ 30 days
Deflection
68% T1
Maintenance
Monthly evals
Compliance
LFPDPPP · SOC 2
06Compliance
LFPDPPPSOC 2Bias evalsPII redaction

HR data is the most sensitive in the company. Provex HR deploys redact PII before it touches any model, run quarterly bias evals across protected categories, and keep every interaction within your HRIS perimeter. LFPDPPP-aligned by default; SOC 2 Type II for the inevitable enterprise vendor review.

07Pack FAQ

Pack-specific questions, answered.

If yours isn't here, the Workshop Offering covers it, or talk to a Forward Engineer directly.

Compensation, conflict, harassment, disability, and termination are guardrailed off the agent. Any question that touches them is routed to a human within the same thread, with the original message intact.
We run quarterly bias evals on every screening agent across protected categories. Results are logged. If drift exceeds threshold, the agent is paused and re-calibrated. No silent degradation.
We integrate with Workday, BambooHR, Personio, Sesame HR, and Runa, plus custom APIs. The HR pack carries adapters out of the box; integration is a Setup deliverable, not a future quarter.
08Get Started

Ready to deploy this pack?

Two ways in. A 2-week Pack Diagnostic, or the Workshop Offering for the upskilling-first path.

Book a call
Typical first deployment ≤ 30 days · SLA-backed maintenance · stack-agnostic · LatAm-compliant