---
title: VRAN-01 — AI Talent Sourcer
slug: vran
role: AI Talent Sourcer
function: Operations
seniority: Senior
verticals:
  - "SaaS"
  - "B2B"
tier: expert
monthlyPrice: $1000/month
publishedAt: "2026-05-04T16:00:00.000Z"
canonical: "https://fidelic.ai/agents/vran"
---

# VRAN-01 — AI Talent Sourcer

*Active sourcing pipelines for hard-to-fill roles*

VRAN runs active sourcing pipelines for hard-to-fill roles — building target lists, drafting first-touch outreach, surfacing candidates against the published role criteria.

## Why it matters

Eightfold and Findem are enterprise-priced (Findem ~$6K/user/yr starting per third-party) with sourcing claims that rarely surface verifiable per-customer outcomes. VRAN is $1,000/mo flat, with the role criteria published and the screening loop auditable.

## Capabilities

- Builds target lists for hard-to-fill roles using public profile data and your published criteria
- Drafts first-touch outreach in your recruiter's approved voice
- Surfaces candidate signals (work history, project authorship, public commits) per the criteria
- Logs every candidate surface with the criterion that matched
- Maintains a per-req sourcing log for the recruiter

## The edge

VRAN refuses to surface or score candidates on protected-class signals. The sourcing loop is auditable per candidate.

## The hook

Every Friday VRAN ships a sourcing-funnel digest: which criteria are matching too few candidates, which req specs may be too narrow, which sources are converting to first-touch replies.

## Tools and integrations

- Slack
- LinkedIn
- Greenhouse
- Lever
- GitHub

## Evidence

- Operates under the four-tier constitution; advance decisions and compensation always escalate
- EvalOps suite covers protected-class signal refusal, criterion-match precision, and false-positive rate
- Inherits new base models and new agent skills automatically — pushed to every Fidelic agent the moment they ship. Same shape as a SaaS update; no upgrade purchase, no version pinning.

## Safeguards

VRAN will not decide which candidates advance. That authority stays with the recruiter and hiring manager.

VRAN will not surface or score on protected-class signals (gender, age, race, national origin, etc.).

VRAN will not commit on compensation; salary discussions escalate to the recruiter.

## Worst-case behavior

VRAN pauses and surfaces ambiguous candidates rather than auto-rejecting. The recruiter decides.

## Day 1 / Week 1 / Month 1

- **Day 1:** Provisioned. Recruiter approves role criteria and outreach voice.
- **Week 1:** First target lists built. First-touch outreach drafted under recruiter review.
- **Month 1:** Per-req sourcing log compounding. Conversion-to-first-reply baseline established.

## 90-day success criterion

Qualified candidates per req per week at the level your head of recruiting defines, with first-touch reply rate above the agreed baseline.

## Integrations / supervision required

ATS integration. LinkedIn Recruiter access. GitHub for technical roles. Slack for digests.

## Resume

**Background.** AI sourcing tools have been recently reframed from 'talent intelligence' to 'agentic'. The actual autonomous behavior in those products is thin. VRAN is built on the Fidelic AI configuration agent with constitutional gates on protected-class screening and advance decisions.

**Methodology.**
- Reads role criteria and recruiter outreach samples before sourcing
- Routes against the four-tier constitution: autonomous on target list building, review-required on first-touch outreach, escalate on advance decisions, refuse on protected-class screening
- Logs every candidate surface with the criterion that matched

**Evals.** EvalOps suite covers protected-class signal refusal (must be 100%), criterion-match precision, and false-positive rate.

**Operating record.** First-cohort deployments scheduled May–June 2026.

## Compatible agents

- [ALEK-01 — AI Chief of Staff](https://fidelic.ai/agents/alek) — Executive coordination that scales
- [OMNA-01 — AI Operations Orchestrator](https://fidelic.ai/agents/omna) — Multi-agent ops coordinator for teams running 3+ agents
- [ZEFA-01 — AI Bookkeeping & Close Lead](https://fidelic.ai/agents/zefa) — Monthly close + reconciliation, with the controller in the loop

## Related honest questions

- [What do I actually own if I cancel my AI agent tomorrow?](https://fidelic.ai/hard-questions/what-do-i-own-if-i-cancel)
- [Is Fidelic just a wrapper around GPT?](https://fidelic.ai/hard-questions/wrapper-around-gpt)
- [Why now? Can't this wait until the category is more proven?](https://fidelic.ai/hard-questions/why-now)

---
Canonical: https://fidelic.ai/agents/vran