Skip to main content
A configured Exterview workflow is deployed when it’s made available for a candidate, recruiter, interviewer, or downstream system to use, as a live agent session or a document workflow. Deploy is the second phase of the Smaya lifecycle. Two things deploy, and it is worth keeping them apart:

Sessions and workflows

An individual unit of work — an interview, a screening, a report. Covered on this page and in the guides beneath it.

AI Employees

A named role (Role Designer, Screener, Coordinator, Report Specialist) you ask for work in Smaya. It asks for your approval before it creates, publishes, or contacts a candidate. See AI Employees and The Hybrid Workforce.
Below you’ll find how deployment works across every agent. We’ve organized this page by the Exterview lifecycle, the agents themselves, the two ways an agent runs, how an agent deploys, and FAQs.

​
The Exterview lifecycle

Deploy is the middle step of the Configure, Deploy, Monitor lifecycle.
  • Configure. Set up your organization structure, evaluation rules, and governance.
  • Deploy. Launch agent sessions and document workflows, and integrate with your systems.
  • Monitor. Review reports, analytics, and post-hire outcome intelligence.

​
Specialized agents, not one model

Exterview is built as a set of specialized agents. Each one owns exactly one session type and holds the full context for that role and candidate, and every agent produces one specific output. There is no single monolithic model trying to do everything. Each session type runs on its own specialized agent that carries the full role and candidate context and produces exactly one report. The full list of agents is below; there is no single monolithic model doing everything.
Exterview runs on Smaya, the enterprise Agent OS. Its agents combine into AI Employees: named roles with a mandate and bounded authority, answering to the people who use them.
One agent, one session type, one output. The agent carries the role and candidate context end to end, which is what makes its output specific and reproducible.

​
The agents

Availability: Available means usable today. Upcoming means not yet available. Upcoming items carry no date.

EX Scan

Available. Resume screening and role creation. Output: Resume Match Report.

EX Call

Available. AI phone screening, including multiple phone rounds per candidate and a mid-call language switch. Output: Phone Call Report.

EX Voice

Available. AI voice interview, on desktop or mobile. Output: Voice AI Screening Report.

EX Avatar

Available. AI video interview with an avatar interviewer. Output: Avatar AI Screening Report.

Panel Interviews in Microsoft Teams

Available. Panel round run in Microsoft Teams, scored from the meeting transcript and recording. Output: Panel AI Report.

HR Round

Available. HR and behavioral round, captured and scored in the same structure as earlier rounds. Output: HR AI Report.

EX Skill Assessment

Upcoming. Role-relevant skill assessment. Output: Assessment AI Report.

EX Code

Upcoming. Coding assessment with structured scoring. Output: Code AI Report.

EX Onboarding Buddy

Upcoming. Post-hire onboarding support. Output: Buddy AI Screening Report.

EX Role Fit

Upcoming. Post-hire role-fit check-in. Output: Rolefit AI Screening Report.
Available today: EX Scan, EX Call, EX Voice, EX Avatar, Panel Interviews in Microsoft Teams, and HR Round. Upcoming: EX Code, EX Skill Assessment, EX Onboarding Buddy, EX Role Fit, and offer management. See Session Types.
Role creation is part of EX Scan and runs as an included document workflow rather than a separate agent. Upcoming: offer management, covering offer recommendation, join-likelihood, the offer document, and e-signature.

​
Two ways an agent runs

Real-time interactions where the candidate engages directly with the agent: phone, voice, and avatar rounds, plus panel and HR rounds in Microsoft Teams. Post-hire companions are upcoming. The agent conducts or captures the session and produces its report on completion.

​
How an agent deploys

1

Enable the agent

Turn on only the agents a role needs, and deploy by role, department, or region.
2

Trigger the session or workflow

A candidate is invited, or a document workflow runs. The agent takes it from there.
3

Agent produces its output

The agent completes and generates its one specific report.
4

Route to review

The output moves into Monitor for human review, with no autonomous adverse decisions.

​
The lifecycle

The deploy lifecycle: a configured agent is deployed, runs as an active session, completes, and its output moves into Monitor for human review.

​
FAQs

Because the agent holds the role and candidate context for that evaluation. That’s what makes its output specific to your role and reproducible over time.
No. Enable only the agents a role needs, and deploy by role, department, or region.
Each agent’s output is detailed in Session Types, and reviewed in Monitor.