NEW
Cadient Talent’s MCP
Connect your ATS to AI.
Learn more →Recruit
Hire
Retain
"Which locations are below headcount?" Ask Cadient Talent's MCP the way you'd ask a recruiter. It answers from live data, inside the assistant your team already uses, and changes nothing.
Sign in with the Cadient Talent credentials you already use. Answers stay scoped to what each person can see today. How your data stays protected →
Trusted by:
100k+ hourly roles filled a year across retail, grocery, healthcare and staffing
Fifteen seconds, start to answer
One question typed in plain English. One answer, built from live data, while you wait.
Illustrative sequence. Real answers come from your own Cadient Talent data.
What is MCP?
The Model Context Protocol gives any AI assistant a secure, structured way into Cadient Talent, so a question becomes an answer, not a report request.
Every answer is read from your ATS as you ask it. Never a stale export.
Ask the way you'd ask a colleague. No query builder, no report menus.
Requisitions, candidates, postings, pipeline, hires: 35 capabilities in all.
Works inside the AI assistants your team already has open.
01 / 05
Funnel counts by stage, and breakdowns by status, location, requisition, source, day or month. This is where the leaks show up.
02 / 05
Look someone up, walk their history, or count and search across a whole intake window: by date, status, location, requisition, posting or source. Onboarding progress included.
03 / 05
Full detail on any requisition, or a one-line summary of all of them: postings, applicant counts by status, hires made, days open.
04 / 05
What's live where, which openings are still unfilled, and which posting each person actually came in through.
05 / 05
Hires by date window, location or source. Average time-to-fill, sliced the way you'd actually present it.
Security
Three things stay true, no matter who is asking or what they ask.
Every answer is read at the moment you ask, and only read. No record is created, edited or deleted. No export, no cached copy, no second system holding your hiring data.
People see through the agent exactly what they'd see in Cadient Talent themselves. Zero permissions are granted that the person didn't already have.
Who asked, what ran, and when. A full audit trail, available to your team and not just to ours.
Also true, and covered in full on the guide: sensitive fields are never returned (government IDs, dates of birth, tax and EEO data, identity documents), credentials are never stored, and access ends the minute someone is deactivated in Cadient Talent.
See the security detail →Getting connected
STEP 01
In your AI assistant, add Cadient Talent as a connector and paste the server address. Nothing is installed.
STEP 02
Your company ID, your Cadient Talent username, your password. No new account is created.
STEP 03
The agent appears in the assistant straight away, already scoped to what you can see.
Supported assistants, step-by-step setup, how access is governed, use cases by role, troubleshooting and the full FAQ.
See it on your own hiring
We'll connect the agent to a live view and let your team ask whatever they'd normally have to file a request for.
Before you ask us
No. Answers are filtered through each person's own Cadient Talent permissions: the same location visibility and the same "my reqs vs. all reqs" scoping they get in the product itself.
Not today. Every capability in this release reads data; none of them create, edit or delete a record. If action-taking becomes available in a later release, it will be governed by the same permissions model.
No. Government IDs, dates of birth, tax data, EEO data and identity documents are blocked at two independent layers and are never returned by any capability.
No. Each person signs in with the Cadient Talent credentials they already use. The password is checked and never stored.
No. Every answer is read at the moment the question is asked. There is no export, no cached copy, and no separate analytics or data-warehouse store.
Claude on web and desktop, the Claude Code CLI, and any MCP-compatible assistant that speaks Streamable HTTP transport. The setup guide lists the current set.
Their access ends when you deactivate them in Cadient Talent. There's no separate user list here to clean up.
Usually one of three things: the hourly / salaried default, a different date window, or "mine vs. all" scoping. Ask the agent to restate the filters it used, then match them to the grid. The setup guide covers both defaults in full.
No. MCP is the standard your assistant uses to reach Cadient Talent. It's the plumbing, not the product.
Still deciding? Book a Session and ask it your own questions instead.
