Login AI
AI recruiter: screening and candidate communications
We set up an AI recruiter for HR teams with 10 or more open vacancies: the agent sorts responses, asks follow-up questions in chat, agrees interviews, and moves statuses into your ATS. A recruiter takes disputed dialogs with the full history.
The AI recruiter reads responses, checks them against vacancy criteria, books interviews, and updates your ATS. You get a short candidate list with the conversation history.
Pricing
from ₽700,000
per project
This is the lower bound of a project. We scope the exact quote after a brief: it depends on volume and the number of integrations.
Capabilities
Screening against vacancy criteria
The agent checks a candidate's answers against the scorecard: experience, stack, location, salary expectations. You get a score for each criterion and quotes from the conversation.
Interview scheduling
The agent offers slots from the recruiter's and interviewer's calendars, confirms the time in chat, and reminds the candidate the day before.
Candidate chat follow-up
The candidate gets an answer about terms, stages, and timing within a minute. The agent writes in Telegram and email, and passes hard questions to a recruiter.
ATS updates
Every step lands in the ATS: status, score, conversation link, and next step. The recruiter opens the card and sees what has already been asked.
From request to result
1
Vacancy and ATS audit
We review 10-20 open vacancies, their criteria, and the response flow. We check which ATS statuses and fields are available through the API and note the bottlenecks.
2
Scorecard and screening rules
We build a scorecard for each vacancy: criteria, weights, and the pass threshold. We write the question rules and the boundaries for the dialog handoff.
3
Integration and pilot
We connect the ATS, calendar, and channels, then launch the agent on two vacancies. For three weeks we compare its scores with recruiter decisions and adjust the thresholds.
4
Launch
We move the whole response flow for the agreed vacancies to the agent. Recruiters work from a short list and take disputed cards themselves.
Check if it fits you
Good fit
A high response volume
Ten or more vacancies bring hundreds of responses a week. The agent sorts them in minutes and gives the recruiter a short list with scores.
An ATS with an open API
Your ATS has an API for cards, statuses, and comments. The agent writes scores and conversation links into the candidate card.
Not a fit
Work without an ATS
If cards live in spreadsheets and email, you need a tracking setup first. We help choose an ATS or build a simple base on PostgreSQL.
Automatic hiring decisions
The final decision and the offer stay with a person. The agent prepares the score and the list, and a recruiter with the hiring manager decides.
What you get
10
10 vacancies in the base
The base setup covers 10 open vacancies and 2 communication channels. That is enough to handle the response flow of a mid-sized HR team and collect funnel statistics.
2
2 channels in the base
In the base configuration the candidate writes in Telegram and by email. Both channels meet in one card, so the recruiter reads the history in one window.
3 wks
3 weeks to the pilot
The audit, integration, and launch on two vacancies take 3 weeks. By then the agent handles real responses, and the recruiter sees the first scores and card handoffs.
What the delivery includes
Scorecard
A scorecard for each vacancy: criteria, weights, pass thresholds, and example answers the agent uses to set a score.
Pilot recruiter
A recruiter who runs the pilot on your side: they check the agent's scores, take disputed cards, and collect edits to the rules.
Pilot report
A report with the funnel by vacancy: how many responses arrived, how many reached an interview, where the agent handed the dialog to a recruiter, and which rules to adjust.
How the system is built
Candidate intake
A candidate writes in Telegram, by email, or in the career page chat, and the agent builds a card right away: name, vacancy, contacts, and a link to the resume. The answer arrives within a minute, and the event goes to the queue and into the ATS.
Scorecard screening
The agent asks questions from the scorecard, checks the answers against the criteria, and sets a score for each item. The card keeps quotes from the conversation, so the recruiter sees the basis for the score and checks a disputed answer fast.
Slots and handoff
When the score passes the threshold, the agent offers calendar slots, confirms the time, and sends a reminder. The recruiter gets a card with the score, quotes, and next step, and the handoff carries the full history.
Where the solution pushes back
- 01
Candidate experience needs setup
Your team sets the tone, response speed, and wording, and the agent keeps them in every dialog. For the first two weeks a recruiter reads the conversations and edits the templates so the candidate gets a clear answer.
- 02
Legal policy and scoring
Your lawyer approves the data rules, retention periods, and field list. The agent writes scores against the scorecard, and a person makes the decision, so disputed cases follow a written protocol.
- 03
ATS data quality
The more accurately you fill ATS fields and statuses, the more useful the report: data quality sets the funnel accuracy. In the audit we show gaps and close them before the pilot.
Technologies
The technologies we use in our work and their role in a project.
Response intake, dialog, and statuses
The agent takes responses through Telegram Bot API, puts events into Kafka, stores cards in PostgreSQL, holds dialog state in Redis, and describes screening rules and model calls in Python.
Cost calculator
This is an indicative range, not a quote. We scope the exact estimate after a brief: it depends on volume and the number of integrations.
Vacancies
Channels
Screening depth
Add-ons
Support
Year one: updates and monitoring, a reply within 2 business days.
Service
AI recruiter: screening and candidate communications
The estimate covers our team's work. Hardware, third-party service licences and travel are added on request.
Estimated cost
₽900,000
Timeline
5 wk
Frequently asked questions
How do you control bias?
The HR team approves the criteria and weights, and the agent scores answers against the scorecard. We remove name, age, and job-unrelated fields from the prompts and test scores on different candidate groups.
Which ATS do you support?
We connect to ATS platforms with an API: Huntflow, SberPodbor, Talantix, and other systems with open methods. For closed systems we set up an exchange through an export or an intermediate database.
Which channels does the agent use?
Telegram, email, and the chat on your career page. The candidate picks a convenient channel, and the conversation history stays in one card.
How long does the pilot take?
The pilot takes 3-4 weeks: the first week covers the audit and integration, the next two cover screening of real responses. By the end you see the share of candidates who reach an interview.
What teams add to this service
Part of Solutions
Implementation Examples
−70%
Request handling time
Ready to discuss your task?
Leave a request – we'll get back with a proposal and a preliminary estimate.
Contact us