> ## Documentation Index
> Fetch the complete documentation index at: https://docs.apollospace.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Recruiting & HR

> Recruiting firms and HR/staffing services: candidate intake, screening scheduling, follow-up, and document collection — with hiring decisions always human.

It's 5pm and the role opened this morning. Forty résumés have already landed
across three channels — an email, a form, a WhatsApp — three candidates from
last week's process have been waiting two days for an answer, the client asked
"how's my search going?" and nobody replied, and the onboarding paperwork for
the one who got hired is scattered across five messages. The recruiter isn't
idle — they're drowning in busywork, not in decisions.

**What leaves your plate:** organizing who comes in, scheduling the screenings,
chasing follow-ups with candidates and clients, and gathering onboarding
paperwork all start happening on their own — with context, within budget, and
with **the hiring decision, the offer, and any judgment about a person always in
a recruiter's hands.**

## How it works, in one line

Recruiting is, at bottom, **relationships and judgment about people** — and
almost everything around it is logistics. The agent doesn't decide who's hired
or send an offer — it takes the friction out of the surroundings (intake,
calendar, chasing, documents) so the recruiter's hours are freed for the
conversation, the read on fit, and the call.

```mermaid theme={null}
flowchart LR
    A[Trigger<br/>résumé arrives, screening booked,<br/>candidate without a reply, approved for onboarding] --> B[Agent pulls context<br/>from Company Brain + CRM + Folders]
    B --> C[Decides + uses tools<br/>CRM, Boards, Routines, Documents, Outbound]
    C --> D{Decision about a person,<br/>offer or contract?}
    D -->|No| E[Organizes / schedules / chases / collects]
    D -->|Yes| F[Proposes + calls the recruiter<br/>with everything ready to decide]
    E --> G[Becomes memory<br/>the next role starts ahead]
    F --> G
```

## Concrete flows

### 1. Intake and organize who comes into the pipeline

**Trigger:** a candidate arrives — a résumé in the inbox, a reply on the role's
form, a referral on [WhatsApp](/en/integrations/whatsapp).

[Athena](/en/agents/athena), the organization's operator, reads what came in,
creates or updates the candidate as a [lead in the CRM](/en/features/crm) on the
right role's pipeline, and logs the intake activity. It attaches the résumé to
the candidate's [folder](/en/features/pastas), extracts the basics — name,
contact, seniority, source — and cross-references the
[Company Brain](/en/features/brain): the role description, the requirements the
client asked for, the candidate's history if they've come through before.
Instead of 40 loose résumés across three channels, you open the pipeline and see
**one organized queue, each at the right stage, no duplicates.**

<Note>
  The agent **organizes and queues** — it doesn't reject, score fit, or decide
  who advances. The read on the candidate and the stage move are the recruiter's;
  the agent just makes sure nobody arrives and gets lost.
</Note>

### 2. Schedule screenings and interviews — with reminders on both sides

**Trigger:** a candidate moves to the screening stage, or the recruiter says
"book the interview with this one."

The agent checks the calendar via
[Google Calendar](/en/integrations/composio), finds the open slots, and proposes
the meeting. Once confirmed, it creates the event, sends the invite and details
to the candidate through [Marcus](/en/agents/marcus) — email or
[WhatsApp](/en/integrations/whatsapp) — and opens a
[task](/en/features/boards) on the board linked to the candidate. The day before,
a [routine](/en/features/rotinas) the agent schedules for itself fires the
**reminder on both sides** — candidate and interviewer — and marks who confirmed
and who didn't. A no-show becomes a visible exception, not a surprise.

<Info>
  **Why this matters:** a lot of candidate drop-off happens in the gap between
  "said yes" and "showed up." A reminder the day before, on the channel the
  person actually uses, holds attendance without the recruiter becoming the
  secretary of their own calendar. The external send can go through approval on
  each dispatch, when you want to review the tone before it leaves.
</Info>

### 3. A follow-up routine for candidates and clients mid-process

**Trigger:** a daily [routine](/en/features/rotinas) — it sweeps open processes
and surfaces who's been stalled without a reply for too long.

On each run, the agent cross-references the [CRM](/en/features/crm) and the
[Company Brain](/en/features/brain): candidates who screened and got no feedback,
clients who received a shortlist and didn't respond, roles with no movement for
days. For each case it opens a [task](/en/features/boards) for the recruiter,
suggests the next step, and — when the content is a routine update ("we're moving
forward with your process, next step next week") — **drafts the message** through
[Marcus](/en/agents/marcus) for you to review and send. A client who gets a
status on time trusts the search is moving; silence is what loses candidates and
contracts.

<Warning>
  Any follow-up that carries **judgment about a person** — an "unfortunately we
  won't be moving forward," a rejection note, any message that communicates a
  decision — is always a draft for the recruiter to review and send. The agent
  **reminds, organizes, and prepares**; the one who communicates a decision about
  someone is a person.
</Warning>

### 4. Collect onboarding documents and keep them organized

**Trigger:** a candidate is hired and enters the onboarding stage.

The agent opens a document checklist on the [board](/en/features/boards) — ID,
proof of address, exams, the signed contract, whatever that client requires —
pulling the template from the [Company Brain](/en/features/brain). As each
document arrives, it files it in the onboarding [folder](/en/features/pastas),
checks the item off, and chases what's missing with a proactive reminder on the
candidate's channel. When the set is complete, it raises a
[notification](/en/features/rotinas) on the bell: "so-and-so's onboarding has
complete paperwork, ready for review." What was a hunt for lost attachments
becomes a closed, checkable folder.

```mermaid theme={null}
flowchart TD
    A[Approved for onboarding] --> B[Agent builds the document<br/>checklist from the Brain]
    B --> C[Chases + files each doc<br/>in the onboarding folder]
    C --> D{Set complete?}
    D -->|Item missing| E[Proactive reminder<br/>on the candidate's channel]
    D -->|Complete| F[Notifies the recruiter<br/>for review and validation]
    E --> C
```

## What stays human

<Warning>
  In recruiting, the decision is about people — and people are always human. The
  agent **proposes and prepares, never decides on its own**, when it comes to:

  * **Hiring or rejecting** — the choice of who advances, who's in, and who's out
    of the process. Always the recruiter's.
  * **Offer and compensation** — salary, package, the terms of the job offer. The
    draft may be the agent's; the decision and the send are yours.
  * **Any judgment about a person** — read on fit, rejection feedback, candidate
    evaluation. The agent doesn't opine about people.
  * **Client contracts and fees** — the placement fee, service terms, any
    commercial commitment.
  * **A message that communicates a decision**, sensitive candidate data,
    governance decisions, and destructive actions.

  In these cases the agent prepares everything — the pipeline, the calendar, the
  draft, the checklist — and **waits for a person's approval**. Every action stays
  in the auditable trail, with author and rationale, isolated per organization and
  LGPD-aligned — which matters, because candidate data is personal data.
</Warning>

## Where to start

<Steps>
  <Step title="Teach your roles and templates to the Brain">
    Upload to the [Company Brain](/en/features/brain) the role descriptions, the
    requirements each client asks for, and the onboarding checklist you use.
    That's where the agent pulls how to queue a candidate and what to chase in the
    paperwork.
  </Step>

  <Step title="Connect calendar and contact channel">
    Link [Google Calendar](/en/integrations/composio) and
    [WhatsApp](/en/integrations/whatsapp) so the agent can propose slots and fire
    the both-sides reminders — always with the option to review the send.
  </Step>

  <Step title="Schedule the follow-up routine">
    Create a [routine](/en/features/rotinas) that sweeps open processes and
    surfaces who's stalled. One agent, one cadence, one channel — and autonomy
    grows like a ratchet, never a leap.
  </Step>
</Steps>

## Next steps

<CardGroup cols={2}>
  <Card title="People & HR" icon="user-group" href="/en/use-cases/people">
    The onboarding, follow-up, and people-organizing pattern applied to your own
    internal team.
  </Card>

  <Card title="CRM" icon="filter" href="/en/features/crm">
    Where each candidate becomes a lead, moves through the role's pipeline, and
    logs activity.
  </Card>

  <Card title="Routines" icon="clock-rotate-left" href="/en/features/rotinas">
    The proactive daemon that chases follow-ups and remembers screenings before a
    candidate goes cold.
  </Card>

  <Card title="Commercial & sales" icon="handshake" href="/en/use-cases/commercial">
    How the relationship with the search's client becomes a pipeline, a proposal,
    and a follow-up.
  </Card>
</CardGroup>