Should I use AI as a recruiter, and how?

Episode one of the Pinpoint How-To Series is done. Here's what you can take from it: five specific things you can do with AI in your recruiting workflow this week, including how to turn a messy kickoff call into a structured hiring brief, and how to walk into every interview already prepared. Plus what to actually do with all of this starting Monday, and three lines I'll never let AI cross.

One idea ties everything in this session together: a structured, consistent process gets you a better hire than not having one. Everything else follows from that. AI, when you use it well, is simply what lets you run that process at scale, for every candidate, without cutting corners.

So, should you use AI in your recruiting? Yes, but it depends entirely on how. Below are five places where it genuinely helps and three lines I won't let it cross.

5 practical ways AI sits in my week

1. Intake and role design

The kickoff call is where every search is won or lost. I record that conversation and AI turns it into a proper hiring brief: the must-haves, the real deal-breakers, the evaluation criteria, a first draft of the scorecard. The hiring manager and I argue about the right thing (the actual definition of good) before the search starts, rather than discovering six weeks in that we'd each pictured a completely different person.

Try it yourself

1. Record the kickoff call
2. Drop the transcript into your AI tool with a prompt to structure it into a hiring brief
3. Edit the output: lock in the must-haves, deal-breakers, and scorecard criteria
4. Get hiring manager sign-off before you start sourcing

2. Candidate prioritization

When 500 people apply in a week, brute-forcing the pile top to bottom means your best applicant might be sitting at number 200 while you grind through the first 199. AI sorts that pile against the criteria we agreed at intake, so the strongest matches rise to the top first.

To be clear, it helps me decide who to look at first, but I still read every shortlisted candidate and make every call myself.

Try it yourself

1. Set your criteria at intake, before applications land
2. Prompt your ATS or AI tool to rank applications against those criteria
3. Review the shortlist yourself — AI surfaces candidates, it doesn't select them
4. You make every advance and reject decision

3. Recruiting content

Job descriptions, job adverts, outreach sequences, candidate communications. All drafted in my tone, in a fraction of the time. If you're not sure where to begin with any of this, begin here. It's the lowest-risk, highest-return corner of the whole thing. Nobody's career gets decided because AI wrote the first draft of a job advert, but you claw back hours every week.

Try it yourself

1. Paste your role brief and a short writing sample into your AI tool
2. Prompt for what you need: job description, advert, outreach sequence, or candidate comms
3. Edit for tone and personalize before anything goes out
4. Save the best prompt and output as a reusable template

4. Reporting and process insight

This one quietly became my favorite. I can ask, in plain English, "which of my roles are off-track, and why?" and get a real answer. It could be sourcing, a hiring manager sitting on feedback, or interview-to-offer conversion. I can also catch when a hiring manager's real bar has drifted from what we agreed at intake. That's a far easier conversation when the evidence is right in front of you.

Try it yourself

1. Ask your AI tool in plain English: "which of my roles are off-track and why?"
2. Follow up to pinpoint the bottleneck: sourcing, feedback delays, or offer stage
3. Prompt for a per-hiring-manager update from live pipeline data
4. Take that data into stakeholder conversations early, before things have already gone wrong

5. Interview preparation

A brief lands in my calendar for each interview I have that day: who the person is, their background, context from any past conversations, and scorecard questions tailored to that specific candidate. Every interviewer on the panel shows up prepared. The admin is done before the day starts.

Try it yourself

1. Set up a scheduled prompt to generate a brief each morning before interviews
2. Format: candidate background, prior conversation context, and tailored scorecard questions
3. Push the prep pack to every interviewer on the panel the night before
4. Score consistently using the criteria you agreed at intake, nothing else

Across all five, none of them replace recruiter judgment; they remove admin and create structure.

3 red lines

1. No auto-rejects. No auto-advances.

AI informs the decision; the human makes it. A person owns every reject and every hire, and is accountable for it. The moment a candidate is auto-rejected by a black box, you have a decision you can't explain to that candidate, can't defend to your organization, and can't defend to a regulator.

2. Don't outsource human judgment.

There are things AI can't assess: motivation, integrity, whether someone will thrive on your team, the potential in someone whose story isn't tidy on paper yet. Some of the best hires I've ever made looked unremarkable on paper and were extraordinary in the room. No model would have flagged any of them; it took a real conversation to find that out.

3. Don't set and forget.

Left unsupervised, AI drifts and can quietly bake in bias at scale before anyone notices. That's the real risk here.

Keep protected characteristics (and their proxies, like graduation year or ZIP code) out of your criteria and prompts. Apply the same bar to everyone. Actually audit the tool for adverse impact. And put a human at each decision point, by design.

Key takeaways

• AI does the work. Humans make the decisions.
• The best AI use cases remove admin and create consistency. They're not the flashy ones, but they're the ones that actually hold a hiring process together.
• Start small. Pick one workflow and improve it. You don't need to transform your whole function overnight, and people who try usually end up trusting none of it.

What to do this week

Three concrete things:

1. Review where AI already touches your process. Is it making any decisions that should still belong to a human? If yes, that's the first thing to fix.
2. Identify the one part of your hiring process that feels most repetitive, manual, or frustrating. That feeling is the signal. Ask where AI could save you time using tools you already have, before buying anything new.
3. Pick one workflow and try it this week. The list above is a good starting point.