XATZHiring, run by agents.
Decided by you.
XATZ ranks the candidates, Zoya calls them, Zia interviews them, and you get a scored shortlist. The waiting drops out, so a shortlist takes days, not weeks.
14 days free. No credit card. Agents included.
The average hire takes over a month. Almost none of it is deciding.
Weeks go to finding people, chasing replies, first calls and collecting notes. That is the work Zoya and Zia can carry, so the calendar collapses to the part that needs you.
Typical pipeline
week 0 of 6Every stage waits on a recruiter's calendar. Candidates go cold in the gaps.
With XATZ
day 0 of 3Zoya works the call list in business hours, Zia interviews when candidates are free. Your decision arrives with the evidence attached.
- Sourcing is manual.
- Recruiters juggle job boards and spreadsheets, then re-key the same candidate three times.
- Screening doesn't scale.
- Every call, every question and every evaluation lands on a person, even at a hundred applicants a role.
- Evidence is scattered.
- Resumes, call notes, recordings and scorecards live in different tools, so decisions get made from memory.
One pipeline. Two agents. Your call.
Zoya handles outreach and Zia handles interviews. AI ranks and scores around them, automations move candidates between the steps, and the last hand-off is to you.
- 1
Find and rank
Search LinkedIn, Naukri, Indeed, Dice and your own database. Every profile is scored against the role.
One ranked list, no duplicates.
- 2
Zoya reaches out
Calls the candidates you put in a campaign, answers their questions about the role, gauges interest, books callbacks and follows up by email.
Interested candidates, outcomes written to the record.
- 3
You screen, with kits
Screening questions generated for the role and a structured evaluation for your notes, so your screening calls are consistent.
Cleared candidates, comparable notes.
- 4
Zia interviews
A live voice interview with a coding task and a behavioural round. Recorded, proctored and scored.
Transcript, score and flags per candidate.
- 5
Report and recommendation
Every interview gets a scored report with a hire / no-hire recommendation and the evidence behind it.
A shortlist you can defend.
- 6
You
Review the shortlist with the transcripts, scores and flags attached. Advance, reject or make the offer.
The hire is yours.
Automations connect the steps: invite interested candidates to Zia, move stages, notify the team. Each automation has an autonomy level, from Off to Decide-and-notify-me, and anything at Suggest, I approve waits in your approval inbox.
Agents inside your pipeline. You set how far they go.
Zoya and Zia do the talking. Automations move candidates between them and bring the exceptions back to you.
- Build the flow once.
- Trigger, AI step, action. Compose it in Flow Studio and it runs every time a candidate reaches that point, for every role you attach it to.
- Set how far it goes.
- Four autonomy levels per automation, from Off to Decide-and-notify-me. The default is Suggest, I approve. Tighten it for a sensitive search, loosen it for volume.
- Anything uncertain waits for you.
- Steps set to ask first land in your approval inbox with the context attached. Approve, edit or reject in one place.
- Every run has a receipt.
- Each call, message, stage move and ranking is logged with what triggered it, so you can audit a run or explain a decision.
Suggest, I approve. The default. Nothing advances or is rejected until you act. Override is one click.
Campaign started: 41 candidates, Senior Backend Engineer.
Called 18 in the morning window. 11 interested, 4 not interested, 3 voicemail.
Callbacks booked for the 23 not reached, inside business hours.
Follow-up email sent to the 11 interested. Outcomes written to each record.
Invite the 11 to a Zia interview? Waiting in your approval inbox.
Design the pipeline. Then let it run.
Every team hires differently. Build the stages your roles need, then attach automations that carry candidates between them.
- TriggerCandidate marked Interested
- AI stepRank against the role, 72% bar
- GateSuggest, I approve
- ActionInvite to Zia interview
- ActionMove to Screen, notify recruiter
- Your stages, per job and per client.
- Drag stages into the order each role needs. An agency can run a different pipeline for every client; an in-house team, for every department.
- Automate any boundary.
- Attach an automation where a candidate crosses from one stage to the next. Trigger, AI step, action, gate. Compose it once, reuse it on every role.
- Agents are steps in the flow.
- Zoya and Zia are nodes you place, not separate tools. A flow can call candidates, invite them to an interview and move them on when the score lands.
- Gates keep you in it.
- Drop a gate anywhere you want a human decision. Everything at that gate waits in your approval inbox with the context attached.
Walk me through how you'd cache this endpoint without serving stale prices.
Write-through for the hot keys, short TTL on the rest, and I'd version the cache key by catalogue release.
Good. Let's make it concrete: implement the key builder in the editor.
Meet Zia. Every candidate gets a real interview.
Zia is the agent that runs the first round. A live voice conversation, technical when it needs to be, and identical in quality for candidate one and candidate one hundred.
- Voice first, distraction free
- A live conversation, not a video call to perform for. Candidates focus on the questions and their code. Nothing else on the screen competes.
- Coding tasks built in
- Candidates write code in the interview. Zia runs it, scores it and adapts the follow-up questions.
- Integrity checked
- Identity, environment and attention signals monitored throughout, flagged where it matters.
- Scored into the record
- Transcript, score and flags land on the candidate the moment the session ends.
Prefer a human round after? Built-in video rooms record into the same candidate record, so the final panel sees everything Zia did.
What changes for your team.
Three parts of the week that stop landing on a recruiter's calendar.
Calls happen without a recruiter dialling.
- Before
- A recruiter works a call list between meetings. Half the candidates never hear back and go cold.
- With XATZ
- Zoya calls everyone in the campaign inside business hours, books the callbacks and writes each outcome to the record.
Everything your ATS did. Included.
The agents need a system of record to work in, so XATZ ships one. Nothing to integrate, nothing to migrate twice.
Jobs and pipelines
Per-client, per-job stages you can rearrange by drag and drop.
One candidate record
Never duplicated, shared across sourcing, screening and interviews.
Sourcing
LinkedIn, Naukri, Indeed, Dice and your database in one search.
Calls and screening kits
Zoya campaigns, call logging, AI screening kits and structured evaluations.
Video rooms
Human-led interviews, recorded and transcribed in place.
Reports
Recruiter performance, drop-off, source effectiveness. Scheduled or on demand.
Candidate portal
Candidates track applications, see interviews and practise before the real thing.
Roles and audit
Granular permissions and a full trail of who did what.
Built for the teams that hire the most.
Staffing agencies
Run every requisition for every client from one place. Zoya keeps candidates warm across roles, and you never lose track of who you placed where.
- Client and contact management
- Per-client interview pipelines
- Submission tracking across jobs
In-house talent teams
High-volume hiring across departments without adding headcount. Each role sees the slice of the pipeline it owns.
- Hiring manager dashboards
- Department-level access
- Native hiring analytics
Your data. Your candidates. Your decisions. Protected.
Agents handle candidate data all day, so the controls around them are not optional. Here is exactly what protects it.
Data security
Your candidates' data is yours alone, and it stays that way.
Every record is scoped to your organisation at the database layer. No query crosses tenants.
Encrypted in transit (TLS 1.2+) and at rest (AES-256). Recordings and documents live in access-controlled object storage.
Role-based permissions down to individual actions. Interviewers see only the candidates they assess.
A complete audit trail: every login, agent action, call, stage move and export, with who and when.
Privacy
Candidates know when AI is involved, and you control what happens to their data.
Candidates are told an interview is AI-led and recorded before it starts, and can decline. Zoya identifies herself as an AI on every call.
Candidate data is used only to run your hiring. It is never sold and never used to train models.
Delete or export a candidate's data on request. Retention periods are set per organisation.
Clear exports for third-party audits, so you can meet local rules on automated hiring tools.
Integrity
Every score is earned by the candidate, and every decision is traceable to a person.
Identity check at the start of every Zia interview, so the person interviewed is the person hired.
Environment and attention monitoring throughout the interview, with the recording analysed afterwards for cheating signals.
Signals are flagged for you, never auto-rejected. You see the evidence and make the call.
Agents never advance, reject or offer. Those actions carry a human name in the log, always.
Simple pricing. Agents on every plan.
Start with 14 days free. Pick a plan when the first shortlist lands.
Free trial
Every agent, every feature. No credit card.
- 2 seats
- 10 Zia interviews
- 100 candidate credits
- 100 outreach minutes
Starter
For a recruiter running their own desk.
- 2 seats
- 10 Zia interviews a month
- 250 candidate credits per seat
- 200 outreach minutes
Growth
For agencies and teams hiring every week.
- 5 seats
- 25 Zia interviews a month
- 500 candidate credits per seat
- 500 outreach minutes
Enterprise
For high-volume hiring across departments.
- Unlimited seats
- Unlimited interviews, credits and outreach
- Dedicated success manager
- Custom SLAs and data residency
One candidate credit is one candidate sourced or imported into XATZ. Outreach minutes are agent call time.
Questions recruiting leaders ask.
Zoya calls the candidates you put in a campaign, answers questions about the role, gauges interest, books callbacks inside business hours and sends follow-up emails, then writes the outcome to each candidate. Zia runs the interview you or an automation schedules, with a coding task and a behavioural round, and returns a scored report with a hire / no-hire recommendation.
Automations have four autonomy levels, set per automation: Off, Show me only, Suggest I approve (the default), and Decide notify me. At the default, nothing advances or is rejected until you act, and the request waits in your approval inbox. Every run is logged with what triggered it.
No. Zia's recommendation is an input. Advancing, rejecting and making an offer are done by a recruiter or hiring manager.
It depends on how fast candidates pick up and book. Because outreach and interviews no longer wait on a recruiter's calendar, a run that used to take weeks can reach a scored shortlist in days. Most teams are sourcing and screening in their first week on XATZ.
Yes. Stages are configured per job and per client by drag and drop, and automations attach to any stage boundary in Flow Studio: trigger, AI step, action and gate nodes. Zoya and Zia are steps you place in the flow.
Yes. Many teams begin by sending candidates they already have to Zia interviews, then adopt sourcing, Zoya and automations as they go.
Data is isolated per organisation, access is controlled by role, and every interview includes explicit candidate consent and integrity checks. See the security overview for details.
Open a role. Let the agents run.
See a full run on one of your live roles, or start a trial and open one yourself.