Modern hiring teams are under pressure to screen candidates quickly without losing fairness or quality. That’s where Zivaro comes in — an AI-powered interview platform that automates first-round interviews using real-time voice agents. Here’s how it works step by step.
The process of conducting interviews with AI is fairly simple and can be understood in a few clear steps. Let's break down how Zivaro operates from setup to delivering a shortlist. Knowing how it works will help you understand how to use it to your advantage.
1. Define Your Role
Everything begins with the hiring team inputting the specifics of the role they are looking for. You provide the AI platform with the job description, requirements, must-have skills, or qualifications. You can also define the seniority level of the role, which will help the AI agents decide on the depth of the interviews and questions to ask.
2. Set Evaluation Criteria and Agent Instructions
Next, for every role the agent will conduct interviews for, you can customize the interview experience using the following things:
-
Topics to be covered during the interviews: This could be technical, behavioral, situational, or any other topic that is relevant to the role. For example, if you are hiring for a software engineer role, you can define topics like: frontend, backend, database, DevOps, or any other topic that is relevant to the role.
-
Evaluation criteria and scoring rubrics: This will help the AI agents score the candidates based on the criteria and rubrics you have defined. For example, if you are hiring for a software engineer role, you can define criteria like: coding skills, problem-solving skills, communication skills, or any other criteria that is relevant to the role.
-
Agent instructions: This will help the AI agents conduct the interviews in a way that is consistent with the criteria and rubrics you have defined.
Here's an example configuration:
Complete Interview Configuration:
| Section | Details |
|---|---|
| Level | Mid-level |
| Custom Instructions | Conduct technical interview for mid-level experienced full stack engineer |
| Topics | Frontend + React , Backend, API Design, Database, DevOps, Containers |
| Criteria - Technical Knowledge | 40% - General technical knowledge of topics |
| Criteria - Problem Solving | 30% - Problem solving ability |
| Criteria - Communication | 30% - Clarity of thought and communication |
| Configuration Purpose | This configuration ensures that every interview is tailored to your specific role requirements while maintaining consistency across all candidates. The level defines seniority, topics cover technical areas, custom instructions provide specific guidance for the AI agent, and evaluation criteria establish scoring rubrics with weights and descriptions. |
3. Invite Candidates to Interview
Once the opening is created based on your configuration, you can invite all the candidates you want to interview using their email addresses. You can send invites to all of them at once. The Zivaro platform will send an invite email to each candidate with the instructions and link to complete the interview.
This saves recruiters hours of scheduling and ensures every candidate gets a fair chance to interview.
4. Candidates Complete the Interview
Once the candidates receive the invite email, they have to accept the invite and complete the interview by joining the live room with the agent. They can complete the interviews at their own time. Questions asked during each interview will be based on the topics, criteria, and rubrics you have defined, as well as the candidates' answers. The agents adapt to the candidates and the interview process.
This flexibility allows candidates to perform at their best while recruiters receive consistent, high-quality data.
5. Agent Scores and Completes Evaluation
During the interview, the agent will regularly score the candidates in the background based on your criteria and review the candidates' answers. The agents adapt to the candidates and the interview process.
The agent understands the candidate's voice, nuances of their speech, and tone. We are not converting the speech to text and then scoring or using text-based AI. The agent receives the candidate's native audio as-is, so the agent understands the confidence, clarity, and tone of the candidate's answers.
This is a key differentiator from other AI interview platforms.
The agent will keep updating the interview summary and feedback for the candidates.
Below is a simple example of the AI interview result:
Complete Interview Evaluation:
| Section | Details |
|---|---|
| Candidate | Ashish - Full Stack Engineer (3-4 years experience) |
| Overall Score | 82/100 |
| Experience Summary | Extensive experience with React and Next.js, including building fully production-ready live applications with active users. Prefers using custom hooks and context providers for state management in React applications. Structured approach to API design for user profile systems. Starts by defining database schema, then designs APIs connected to authentication, including endpoints for creating users, updating profiles, and fetching user profiles. Prefers running database migrations on replica set first to verify structural changes before applying to production. Sets up containerized environments using Docker and Docker Compose with specific Linux versions and coordinated services. Suggests extensive logging on backend to track application issues and toast notifications on frontend for user alerts. Emphasizes sending backend notifications or logging frontend errors on backend for end-to-end error handling. |
| Technical Knowledge | 85/100 - General technical knowledge of topics |
| Communication | 75/100 - Clarity of thought and communication |
| Problem Solving | 87/100 - Problem solving ability |
| Detailed Feedback | Demonstrated strong problem-solving skills, especially in designing safe database migration strategies and setting up comprehensive logging for both frontend and backend. Showed solid technical knowledge in containerization with Docker and Docker Compose. Communication was clear most of the time, though a bit more structure in responses could help enhance clarity. Overall, excellent understanding of full-stack engineering practices, with minor areas for improvement in organizing thoughts during explanations. |
How Zivaro Differs from Other AI Interview Platforms
Unlike most AI interview tools that convert speech to text and analyze transcripts, Zivaro processes candidates’ native audio directly — understanding tone, confidence, and clarity in real time. This allows for far more human-like and accurate evaluations. It ensures recruiters get a deep, context-rich understanding of each candidate’s communication and confidence — something text-only systems simply can’t deliver.
6. And You Get a Shortlist
As the interviews are getting completed, the platform will keep the shortlist ready for all the interviews conducted for your job opening. You can review for any given role all the candidates who have completed the interviews. You will have access to their summaries, feedback, scores, and the overall score. You can also export the shortlist to your ATS or email it to your hiring team.
To summarize the process: calibrate the AI with what you need, instruct the agent to conduct the interviews, and get a shortlist of candidates who have completed the interviews. It's a seamless pipeline from job description to shortlisted candidates. Depending on how fast your candidates complete the interviews, you can create a job opening and have actionable insights on candidates for next rounds within hours.
Start Hiring Smarter with Zivaro
Don’t let great candidates slip through resume filters.
Zivaro’s AI interview platform automates candidate screening, evaluates real skills, and delivers ranked shortlists in record time.
✅ 80% faster screening
✅ Fair, bias-free interviews
✅ Consistent scoring across all roles
👉 Book your demo now and transform your hiring workflow with AI.