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

# Smaya Assistant

> Ask Smaya about your own hiring data and get cited answers, streamed as they are written, with a view of what the agent did. Direct your AI Employees, and chat with any candidate report.

<img src="https://mintcdn.com/exterviewinc/2F3WHDYe4q7V4uTB/images/smaya/smaya-mascot.webp?fit=max&auto=format&n=2F3WHDYe4q7V4uTB&q=85&s=d2c746965d5bc4073d6ed5a472668797" alt="Smaya" width="96" height="96" className="smaya-hero smaya-hero-sm" data-path="images/smaya/smaya-mascot.webp" />

## Purpose

The **Smaya Assistant** is the conversational surface of the platform. It answers from your organization's own data with citations, and it is how you direct the work of your [AI Employees](/smaya/ai-employees).

It is one surface of the OS, not the whole of it. The substance is the AI Employees doing the work; the assistant is how you ask about that work and direct it.

<Info>
  **Status: Available.** Ask Smaya, streamed answers with a view of what the
  agent did, report chat, memory across conversations, and actions that wait for
  your approval are all Available.
</Info>

## Where you can ask

<CardGroup cols={2}>
  <Card title="The Smaya page" icon="message-square">
    **Available.** Open **Smaya** in Exterview to ask anything across your workspace — jobs, candidates, interviews, reports and rounds.
  </Card>

  <Card title="Beside a candidate report" icon="file-text">
    **Available.** Report chat opens Smaya on the candidate report you are reading
    and answers about that report, with citations to the evidence.
  </Card>

  <Card title="Your own assistant" icon="plug">
    **Available.** Claude, ChatGPT and Microsoft 365 Copilot reach the same AI
    Employees over MCP. See [Microsoft & Interoperability](/smaya/mcp).
  </Card>

  <Card title="Slack" icon="hash">
    **Upcoming.** Ask an Exterview AI Employee in a direct message or with an @mention. Slack will be read-only: nothing is changed from Slack.
  </Card>
</CardGroup>

## Streamed answers, and what the agent did

**Available.** Answers stream in as they are written, so you are not waiting on a blank screen. A transparency panel above the answer shows what the agent did to reach it: which AI Employee handled the question and which reads or actions it ran.

The panel says what the agent did. The citations say where each point came from. Together they let you check an answer without taking it on trust.

## Ask and understand

Read-only. Nothing changes.

<CardGroup cols={2}>
  <Card title="Understand a person's evaluation" icon="user">
    A plain-language summary of fit, strengths and gaps against the role.

    *"Summarize this candidate's fit for the role."*
  </Card>

  <Card title="Read back a report" icon="file-text">
    The key points from any completed evaluation, without opening and scrolling it.

    *"What did the phone screen flag?"*
  </Card>

  <Card title="Compare" icon="scale">
    Finalists side by side on the dimensions that matter.

    *"Compare the top three for this role."*
  </Card>

  <Card title="Explain a result" icon="chart-line">
    The reasoning and the evidence behind a score, so a decision is defensible.

    *"Why did this candidate score well on communication?"*
  </Card>

  <Card title="Check status" icon="workflow">
    Where a role stands, who is waiting, and on whom.

    *"Who is waiting on a decision for this job?"*
  </Card>

  <Card title="Pick up where you left off" icon="history">
    Smaya keeps your conversation history and can recall earlier conversations.

    *"What did we decide about the analyst shortlist last week?"*
  </Card>
</CardGroup>

## Act, with your approval

**Available.** When you ask for a change, Smaya hands the work to the AI Employee whose job it is. Creating a role, publishing a job and contacting a candidate always show an approval card first.

<CardGroup cols={2}>
  <Card title="Draft and publish a role" icon="file-plus">
    Role Designer turns a brief into a drafted role with questions and scoring, and publishes it once you approve.
  </Card>

  <Card title="Invite candidates to screening" icon="send">
    Screener reads applicants against the criteria and proposes who to invite.
  </Card>

  <Card title="Schedule panel rounds" icon="calendar">
    Coordinator schedules, reschedules and cancels rounds, and refuses a real
    clash rather than double-booking.
  </Card>

  <Card title="Start a role run" icon="play">
    **Upcoming.** From a connected assistant, start a run for a role. See [The Agent Swarm](/smaya/agent-swarm).
  </Card>
</CardGroup>

## The line that does not move

<CardGroup cols={2}>
  <Card title="Grounded and cited" icon="quote">
    Every answer traces to a source. When it does not know, it says so rather than inventing an answer.
  </Card>

  <Card title="Scoped to you" icon="shield-check">
    It reaches only what you could reach yourself, within your organization and
    your own permissions.
  </Card>

  <Card title="Asks before it writes" icon="user-check">
    Consequential changes wait on your approval. No hiring outcome is ever set by
    the assistant.
  </Card>

  <Card title="Fully recorded" icon="clipboard-list">
    Every action and approval is captured and attributed.
  </Card>
</CardGroup>

## FAQs

<AccordionGroup>
  <Accordion title="Is the assistant the same thing as Smaya?">
    Smaya is the Agent OS, and also the name of the general-purpose AI Employee that answers across your workspace. The assistant is its conversational surface. When someone says "ask Smaya," they mean this.
  </Accordion>

  <Accordion title="Do I need specific phrasing?">
    No. The prompts here are illustrative. It interprets natural requests in the
    context of what you are viewing.
  </Accordion>

  <Accordion title="Can it do something I am not allowed to do?">
    No. It operates entirely within your own role and permissions.
  </Accordion>

  <Accordion title="Can I ask from Slack?">
    Not yet. Asking from Slack is Upcoming, and it will be read-only. Today, changes are made in Exterview or from a connected assistant, with approval.
  </Accordion>
</AccordionGroup>

## Related

<div className="kb-links">
  *   <span className="kb-ico">
        <Icon icon="id-card" />
      </span>
    [AI Employees](/smaya/ai-employees). The workforce the assistant helps you
    direct.
  *   <span className="kb-ico">
        <Icon icon="message-square" />
      </span>
    [How to use Smaya](/guides/smaya-companion). Step-by-step in the app.
  *   <span className="kb-ico">
        <Icon icon="shield-check" />
      </span>
    [Governance & Control](/smaya/oversight). The approval model.
</div>