Skip to main content

​
Purpose

Understand the vocabulary and mental model before you go deeper. Every other page assumes these terms.

​
Key objects

Organization

Your isolated organization environment. All data, configuration, and scoring are scoped to it.

Role

A job definition. The structure a candidate is evaluated against.

Candidate

A person being evaluated for a role.

Evaluation

A scored assessment of a candidate against a role’s rubric.

Agent Session

A live, real-time interaction (phone, voice, avatar, panel or HR) run by an Exterview agent. Assessment and post-hire sessions are Upcoming.

Document Workflow

A non-live task such as role creation or resume match. Offer generation is Upcoming.

Report / Takeaway

The structured output every completed session or workflow produces.

Rubric

The set of dimensions a candidate is scored against.

Configuration Contract

The signed, versioned, frozen record of all setup decisions that every evaluation score traces back to.

Human Review

A person reviews and approves outcomes; nothing adverse is decided autonomously.

Outcome Intelligence

Post-hire signals (D30, D60, D90) fed back to improve evaluation quality.

​
Live sessions vs. document workflows

Exterview runs two kinds of work:
  • Live session. A real-time interaction: phone, voice, avatar, panel and HR. Upcoming: assessment and post-hire sessions.
  • Document workflow. A non-live task: role creation and resume match. Upcoming: offer generation.

​
AI and Core capabilities

Every capability in Exterview is one of two kinds, and the guides label each page accordingly:
  • AI. Powered by Exterview’s models: screening, interviewing, matching, and scoring.
  • Core. Standard platform work: jobs, candidates, users and roles, billing, integrations, and configuration.
The split matters because AI capabilities produce scored evidence a human then reviews, while Core capabilities are the platform tasks that set them up and run around them. See the guides for the labelled how-to pages.

​
Agents and AI Employees

Exterview runs on agents: specialized workers, each accountable for one job. You don’t build them; they ship pre-built, and you configure the criteria and scope they work to. See Exterview Agents. Agents combine into AI Employees: named roles for AI on your team. Four are available today (Role Designer, Screener, Coordinator and Report Specialist), plus Smaya as the general-purpose assistant. They run with the access rights of the person using them, and creating a role, publishing a job or contacting a candidate always waits for approval. Your teams work with AI Employees, not with raw agents. See AI Employees. Both sit on Smaya, the enterprise Agent OS that provides governance, integration, evidence and lifecycle to every application above it. See Smaya.

​
Independent review of every output

Available today. Automated checks run on every scored output (schema, score range and citations), and the results show in Scoring Logs under Settings. Every score records the model and prompt that produced it. Upcoming. A Quality & Fairness Agent that independently reviews every agent’s output before it becomes part of the record. See AI Observability and Governance & Control.

​
The agent swarm

The agent swarm is Upcoming. As designed, you give Smaya a role, and it builds the pipeline, talks to every candidate, gathers the evidence and brings you the people worth deciding on. You see outcomes and reports, not agents.
  • Available today. Screen and rank candidates through EX Call, EX Voice and EX Avatar into the Leaderboard, ask Smaya about any job, candidate or report, connect Claude, ChatGPT or Microsoft 365 Copilot over MCP, and work with AI Employees that ask for approval before they create, publish or contact a candidate.
  • Upcoming. Starting a role run, the Agent Inbox with reports to view (Proceed, Drop, Talk) and schedules to accept (Accept, Decline, Talk), the full role-to-decision run across all five report moments, outreach sequences, standing 24/7 operation, and run-level budgets and pause controls.
No AI system sets a candidate to hired, rejected or withdrawn. Every decision change is recorded with a person’s reason. See Agent swarm.

​
Setting up: the guided journey

Enterprise setup is a guided, phased configuration led by your Customer Success team. You define org structure and what a strong candidate looks like, design the pipeline, set scoring and gates, connect systems, calibrate per domain, and set governance. Every decision is then signed and frozen into the Configuration Contract before go-live. See Enterprise Setup for the full step-by-step path.

​
Configuration tiers

How much you configure depends on your plan. Exterview is designed to land small and expand. See Configure for the full model.

​
Hyper-calibration

For Enterprise customers, Exterview builds a per-organization hyper-calibration layer during Foundation setup. Your competencies, role definitions, and scoring context are captured once, taken through a staging journey, and then frozen, so every evaluation is scored the same way and stays reproducible. This calibration is unique to your organization. Upcoming: post-hire outcomes feeding back in, so it compounds over time. See Calibration and Outcome Intelligence.

​
The evaluation lifecycle

Every evaluation moves through five states: The five states every evaluation moves through, from frozen configuration to a monitored result. Each state is defined in the table below.

​
Governance, by design

Human Review

People approve outcomes before they take effect. Roles are set by the nine system roles.

Auditability

Every action and result is logged in a traceable record.

Data privacy

Data is scoped to your organization, with retention and consent controls.

Responsible AI

Content-based scoring, explainability, and no autonomous adverse decisions.

​
FAQs

A session is live and interactive (an interview). A workflow is a document task, such as creating a role or matching a resume.
AI features are powered by Exterview’s models and produce scored evidence: screening, interviewing, matching, and scoring. Core features are standard platform tasks such as jobs, users, billing, and configuration. The guides label each page.
No. Scoring is content-based, meaning what a candidate says, not biometrics, facial analysis, or tone.
Configuration is frozen per version in the Configuration Contract. When rules change through change control, scoring re-baselines to a new version so results stay reproducible.
Access is role-based and scoped to your organization. See Trust & Compliance for details.
  • Modules. The full catalog of Exterview modules and what each produces.
  • AI Employees. How agents combine into a chartered, governed role.
  • Glossary. Every term, defined.
  • Enterprise Setup. The guided journey to a frozen Configuration Contract.
  • Why Exterview. The category, the moat, and how we’re different.