AI and Recruiters: What's Actually Happening and What to Do

The honest assessment

Recruitment is one of those professions that was already changing before AI showed up, and AI just hit the accelerator. Hard.

Here's what AI can do in recruitment right now. Write job descriptions in seconds. Screen CVs and rank candidates against criteria. Generate boolean search strings for sourcing. Draft outreach messages for passive candidates. Schedule interviews. Produce interview question sets tailored to specific roles. Summarise candidate feedback. Write rejection emails that don't sound like they were generated by a particularly cold robot (though sometimes they still do). Tools like LinkedIn Recruiter's AI features, HireVue, and various ATS platforms have been building AI capabilities for years. The generative AI wave just made everything faster and more capable.

What AI can't do is read people. It can't tell that the candidate who looks average on paper is actually brilliant because she's been stuck in a company that wouldn't promote her. It can't sense that the hiring manager says they want "a self-starter" but actually wants someone compliant who won't challenge them. It can't build the relationships that mean a candidate calls you first when they're thinking of moving, six months before they update their LinkedIn. It can't close. The human persuasion, the trust, the instinct for timing... that's not in the algorithm.

The part that's changing fastest is sourcing and screening. The top of the funnel. LinkedIn's AI can now suggest candidates, generate InMail messages, and predict response likelihood. ChatGPT can create entire sourcing strategies for niche roles. The volume work that used to keep junior recruiters busy... posting ads, reviewing initial applications, scheduling calls... is being automated. If your value is measured in CVs reviewed per day, the machines review more. A lot more.

Your exposure level: High

High exposure. Recruitment is a volume business for many practitioners, and volume tasks are exactly what AI automates.

LinkedIn's own data shows that AI-assisted recruiters can source candidates 2-3x faster than those working manually. That's not a marginal improvement. It's the kind of efficiency gain that changes headcount decisions. If one recruiter with AI tools can do the sourcing work of three, agencies don't need three. They need one plus a subscription. Some of the larger recruitment firms have already reduced team sizes on the back of AI implementation. They just don't tend to announce it publicly.

The agency model in particular is under pressure. Clients are asking why they should pay 20% of a salary for a placement when they can use AI-powered platforms to do much of the sourcing themselves. Indeed, LinkedIn, and specialised platforms are all adding AI features that reduce the need for intermediaries. The agency recruiters who survive will be the ones operating in complex, high-value markets where relationship and judgement matter more than search capability. Executive search. Niche technical roles. Senior leadership positions.

For in-house recruiters, the picture is slightly different. The role is shifting from "find candidates" to "attract, assess, and close candidates while managing the hiring manager relationship." AI handles the first part increasingly well. The rest is still yours. But "the rest" is a smaller portion of time, which means companies may need fewer in-house recruiters to cover the same hiring volume.

The 90-day action plan

1. This week: generate job descriptions with AI. Take your last five job descriptions. Give ChatGPT the role title, key requirements, and company context. Compare the AI output to what you wrote. Chances are the AI version is at least as good, possibly better at inclusive language and SEO optimisation. Painful. But now you know.

2. Week two: automate your outreach. Draft 10 different InMail templates using ChatGPT for different candidate types and seniority levels. Personalise each with specific variables (company name, mutual connection, relevant experience). Test response rates against your standard messages. Most recruiters find the AI-assisted messages perform comparably or better. Use the time you save on the conversations that actually matter.

3. By day 30: build an AI-powered screening process. Paste a job specification into Claude along with five anonymised CVs. Ask it to rank them and explain its reasoning. Compare to your own assessment. Understand where it agrees with you and where it differs. This isn't about trusting AI blindly. It's about using it as a second opinion that works at speed.

4. By day 45: develop interview intelligence. Use Claude to generate role-specific interview questions that go beyond the standard competency framework. Ask it to create scoring criteria. Use it to summarise candidate assessment notes and flag inconsistencies across interviewer feedback. This level of rigour makes your process better, not just faster.

5. By day 60: become a hiring manager whisperer. This is where AI can't follow you. Get better at the intake meeting. Learn to challenge vague requirements. Push back on unrealistic salary expectations with market data (Perplexity is excellent for pulling quick salary benchmarks). The recruiter who improves the quality of the brief improves the quality of every hire. That's value.

6. By day 75: track and quantify your impact. Start measuring things beyond placements. Time-to-fill improvement. Quality of hire metrics. Offer acceptance rates. Hiring manager satisfaction. Use these metrics to tell a story about your value that goes beyond "I filled the role." AI can fill roles. You improve outcomes.

7. By day 90: decide your niche. Generalist recruitment is the area most exposed to AI disruption. Specialist recruitment in complex markets is far more defensible. If you haven't already, pick an industry, function, or seniority level and go deep. Become the person who doesn't just find candidates but understands the market so well that your insights are worth paying for.

AI tools you should be using this week

• ChatGPT for Work — Use it for job descriptions, outreach messages, interview questions, and candidate summary reports. Also excellent for generating boolean search strings for sourcing. Give it the role spec and ask for creative sourcing strategies you haven't considered.

• Claude for Work — Better for candidate assessment work. Paste in a role spec and multiple CVs for comparative analysis. Also good for creating structured interview frameworks and writing thoughtful rejection feedback. Handles nuance better than ChatGPT for sensitive communications.

• Microsoft Copilot for Work — If your ATS integrates with Microsoft 365, Copilot can summarise email threads with candidates, draft follow-ups, and manage the administrative overhead of the recruitment process. The Outlook and Teams integration saves significant time on scheduling and communication.

• Perplexity for Research — Quick salary benchmarking, competitor analysis, company research, and market mapping. When a hiring manager asks "what's the market rate for this role in Manchester?" you can answer in 30 seconds with cited sources instead of promising to check and getting back to them.

What to say in meetings

When a hiring manager says "can't we just use AI to find candidates?": "We can use AI for sourcing, and I do. It's fast. But finding candidates is the easy part. Knowing which ones will actually accept, stay, and succeed in your team is the hard part. That's where I earn my keep." Specific. Confident. Not defensive.

In team meetings about AI adoption: "I've integrated AI into my sourcing and screening workflow. My time-to-shortlist is down by 40%. I'm reinvesting that time in deeper candidate assessment and hiring manager partnership. Happy to walk the team through my process."

When a client or manager questions recruitment fees: "The AI tools are excellent at finding people. What they can't do is persuade a passive candidate to take a meeting, assess cultural fit, manage counter-offers, and ensure someone doesn't fall out of the process at the last minute. That's what you're paying for."

If the worst happens

If you're made redundant from a recruitment role, your transferable skills are genuine and underrated. You can sell. You can persuade. You can assess people quickly and accurately. You can manage multiple stakeholders with competing interests. You can work under pressure with targets. Those skills map directly to sales, business development, talent management, HR, and client relationship management.

The most natural adjacent moves: talent acquisition lead (in-house, focusing on strategy rather than volume), HR business partner, sales or account management, headhunting in a niche market, or recruitment technology consulting. There's actually a growing market for people who understand recruitment and can help companies implement AI recruitment tools properly. Most companies are adopting these tools badly. Someone who knows both the technology and the human side of hiring is rare and valuable.

Final thought from someone who's been through the other side. Recruitment teaches you resilience. You've been ghosted by candidates, screamed at by hiring managers, and missed targets that weren't achievable in the first place. If you can handle recruitment, you can handle career change. Apply the same persistence to your own career that you'd apply to finding a needle-in-a-haystack candidate. You already know how to do this. You just haven't done it for yourself before.