Enterprise · Recruiting
AI recruiting agent
Applicant FAQs from the job doc you approved, structured screening into a table, packets to hiring managers. Humans still decide who gets an offer.
Applicant FAQs from the job doc you approved, structured screening into a table, packets to hiring managers. Humans still decide who gets an offer.
Applicants ask the same five questions. Coordinators paste the same answers. Managers still miss the strong packet because it lived in email.
A recruiting agent faces candidates in customer-facing mode, reads the role doc, writes screening fields, and escalates humans for interviews. It should not silently reject people unless you built and accepted that rule.
Doc
Job as source of truth
Answers from the req you maintain
Fields
Screening you can sort
Must-haves as columns, not prose
Human
Offers stay human
Default is collect and route
Traditional tools
Careers mailbox + ATS + a FAQ chatbot that cannot see the req.
With Maicos
Job doc, form, candidates table, and agent in one hiring workspace.
Consistent FAQs and structured packets. Not an unattended firing squad.
Location, visa, salary band — only what you put in the job doc or table. If it is not written, the agent says it will check with a human.
Must-have questions become fields. A 12-email thread is how details get lost before the manager ever looks.
Above your threshold, @mention the hiring manager with the row, files, and answers attached.
Once a human says yes, the booking agent offers times. The recruiting agent should not invent interviewer availability.
Write the req. Store answers. Keep offers with people. Review any auto-reject rule like it will be read in court — because it might.
The agent will quote it. Vague “culture fit” prompts are how bias sneaks in. Prefer observable must-haves.
If you cannot sort it, you cannot run a fair process at volume.
Auto-reject is optional and should be a documented rule (missing license, cannot work in-region) — not a vibe from the model.
Customer-facing mode. Résumés in the hiring project, not in a random shared drive.
If the req says no, say no. If it is silent, do not guess.
Candidate asks on the careers agent
No employee login.
Agent reads the job doc
Visa field is “not offered.” Answer that, cite the doc.
If they still apply
Form fields land in the table. Manager can still look; you did not hide the constraint.
Strong must-have match
Packet to the hiring manager. Booking only after they accept the screen.
Consistency is the product. A model that “sounds encouraging” about visas you do not offer is a liability.
Candidates deserve the same answer at 2am as at 2pm. That only happens if the source of truth is a document, not the model’s mood.
The same apps you already work in — not a second chatbot that needs uploads.
Useful agents have a job description. These limits keep this page honest — and help you pick a different agent when you need one.
How people use this every day.
Quote the official job doc. Unknowns go to a coordinator, not a guess.
Must-have fields in a table; managers open packets, not inboxes.
A Maicos Quiz for a defined skill, then a score column. Still not an offer.
Different job docs. Do not let one agent blur employment type.
A candidate-facing hiring assistant that answers from your job materials, stores screening data in tables, and hands qualified packets to hiring managers. Offers stay human.
Only if you build an explicit rule. The default described here is collect, score, and escalate.
Files and form fields in the hiring project, next to the candidates table.
You are responsible for local employment law. This page is product capability, not legal advice. Review auto-decisions with counsel.
They can, but internal mobility often belongs on a different form with different visibility. Do not mix internal notes into a public candidate transcript.
Start free — Workspace, Create, Automate, and Integrations built for AI from day one.
No credit card required · Free during beta