5 of 8 places openApplications close 17 October
Your hiring team — more hires, less busywork.
Eight teams. One metric each. Three months. Cara, hired in to move it, in the tools and on the AI model you already run.
However fast you're hiring, you're being asked to do more, per person.
The plan always outpaces the people, whether the team is growing, flat or shrinking. And the first thing to go is never the hire. It's the work around the hire: the intake questions nobody has time to ask, the debrief that gets skipped, the pipeline that never gets followed up. Not because anyone's careless. Because the middle of the week is where the middle of the work goes.
Meet the hire who takes the middle.
Cara by SocialTalent is the AI hire who takes on the transactional work of recruiting, using the Apps and AI model you already run.
She doesn't assist. She does the work. She starts on one Skill, at the approval level you set, and takes on more as she earns it. And she works where your team already works, in Slack or Teams, so there's nothing to log in to and nothing to roll out.
This pilot is where we prove it, against one number you already care about.
Debrief for the Senior Recruiter role is consolidated — three of four panellists in. I've nudged the fourth.
One metric.
Three months.
Your own tools.
We're taking eight hiring teams through a three-month Cara pilot. Each team picks one metric. We agree how it's measured today, hire Cara in on the Skill that moves it, and read the number out at the end. That's the whole design.
Why one metric, not a dashboard. One is honest. A dashboard lets everyone find a number that moved. One metric means we agree up front what "worked" looks like, and if Cara didn't move it, you'll know. So will we.
How the three months run
- Weeks 1–2
Scoping
Your metric, and how it's measured today. The Skill Cara starts on. Who owns it on your side. You set the approval level, from approval-required through to fully autonomous, and autonomous never means unsupervised.
- Weeks 3–12
Cara works
In your Slack or Teams, on your own LLM, inside your own tools. None of your data sits on a SocialTalent server. A weekly check-in with your SocialTalent lead, and she takes on more as she earns it.
- Week 12
The readout
Day one versus day ninety, on the metric you chose. A readout you can put in front of a CFO. What you do next is your call.
Five to choose from. You get one.
- 01
Time to schedule interviews
Req open to interviews on the calendar. The scheduling back-and-forth, gone.
Speed - 02
Time to shortlist
Applications in to a screened list out. Cara screens and structures; your team decides.
Speed - 03
Intake completion
The share of reqs that start with a proper intake, not a job title and a hope.
Completeness - 04
Debrief completion
The share of interviews that end with feedback actually in, consolidated and ready to decide on.
Completeness - 05
Recruiter hours returned
Hours per recruiter, per week, taken off the transactional work.
Capacity
Only two of the five are speed metrics. Two are completeness metrics: the work that should happen on every hire, and doesn't when the week gets busy. One is capacity. That mix is deliberate.
Eight teams. Here's what we're looking for.
It's a paid pilot, scaled to your hiring volume. We'll cover the numbers on the scoping call, not on a web page.
Apply for a place→- 011,000 or more hires a year.The pilot is sized for that volume. If you're under it, apply anyway and we'll talk about the right way in.
- 02Live reqs and a plan that's outpacing the people.Cara needs real work to do from week three.
- 03Slack or Teams in daily use.That's where she works. There is nowhere else to log in to.
- 04An LLM in production, or ready to be.She runs on the AI model you already run. If you're not there yet, tell us on the form and we'll talk it through on the scoping call.
- 05One owner on your side.Someone who can pick the metric, set the approval level and make a call in a weekly check-in.
Proven. Plugged in. Configurable.
The answer to a three-vendor shortlist — the self-build, the new platform and the native ATS feature — in one line.
Trained, never generic
Trained by us, on fifteen years of our own hiring content, not a generic model. She remembers her training and is onboarded to your processes and preferences.
Runs on your stack
She uses your network, your tools and your own LLM. No vendor technology of her own to bring in.
None of your data sits on a SocialTalent server
Your data passes through our infrastructure for a moment, to connect your own systems to your own AI model. It is never retained and never logged, and no one at SocialTalent ever sees it. We see which Skill fired and how often. That's usage, not content.
You set the approval level
From approval-required through to fully autonomous. Autonomous never means unsupervised: Cara can own completion of a task, but a person always owns the employment outcome.
The questions pilot teams ask
It can mean that, or your existing team doing more with the same headcount. That's your call — and we won't pretend otherwise.
Cara can own completion of a task, but people own the employment outcome — she never makes the hiring decision. You set the approval level, from approval-required through to fully autonomous, and autonomous never means unsupervised.
No. She works where your team already works, in Slack or Teams, with nothing new to learn. She runs on your own LLM, inside your own tools, and none of your data sits on a SocialTalent server.
Proven, plugged in, configurable — nobody else answers the self-build, the new-platform and the native-feature question at once.
No. It's a paid pilot, scaled to your hiring volume. We'll cover the numbers on the scoping call.
You get the readout: day one versus day ninety on your metric. If she moved it, most teams keep her on and add a Skill. If she didn't, you'll have the numbers that say so. Either way, the decision is yours.
Tell us on the form. It isn't a rejection. It's the first thing we'll work through on the scoping call.
No. Many of the teams applying are, and they'll recognise the method behind Cara's work. It isn't a requirement.
You get the two ends back.
Apply, and at scoping we'll put your reqs and your one metric on the board — and show you exactly where the hours come back, and exactly where the judgment goes back to your team.
You're in the queue.
We'll come back within two working days to set up a scoping call. In the meantime, the 80 tasks of hiring shows you where your team's hours actually go — and where the pilot would start.
The 80 tasks of hiring→Thanks — this one isn't the right fit yet.
The pilot is sized for teams hiring 1,000 or more a year. We'll be in touch about the right way to bring Cara in for your team. In the meantime, the 80 tasks of hiring shows where your team's hours actually go.
The 80 tasks of hiring→Cara by SocialTalent is the AI hire who takes on the transactional work of recruiting, using the Apps and AI model a customer already runs. Built on fifteen years of hiring best practice, she's brought in one skill at a time — alongside SocialTalent's Learning platform for teams who want to elevate the people doing the work that remains. Your hiring partner for the AI era.