AI is already part of recruiting (whether we admit it or not). But how do you put AI in recruitment to work without losing the human touch?
Nobody wants a flood of generic outreach clogging their LinkedIn messages. And no TA leader wants to explain to their hiring manager why the ATS just auto-screened out the strongest candidate in the pile.
To find out more, we spoke with Joe Atkinson, Scede’s Director of AI and Automation, to understand how recruiters are currently using AI, the benefits and challenges it presents, and what integrating AI into your day might look like.
The state of AI in recruitment in 2025
AI is creeping into the tasks recruiters perform every day. It doesn’t always feel dramatic in the moment, but step back and the scale of change is hard to ignore. As Joe says, “The first wave has been admin. Note takers like Metaview, Equitas, even Gemini or Granola. People are comfortable with them because they’re generally low risk. They record and summarise, but don’t make decisions.”
Likewise, outreach followed. “AI is writing a lot of first messages right now,” Joe said. “But most of it isn’t great. You can spot it immediately. The idea of personalisation at scale is strong, but the execution still needs a human touch to feel authentic.”
Interestingly, sourcing is starting to change, too. “It’s early days, but we’re actively looking at how recruiters can effectively use AI for sourcing.”
Screening and interviewing are moving more slowly, though there are experiments. Joe said, “I’ve seen voice AI used for first-stage interviews, and interestingly, candidates actually talk more to an AI. They give longer answers, maybe because they feel less judged. That surprised me.”
In fact, a Gallup survey found that 93% of Fortune 500 CHROs have already begun integrating AI tools into their people functions. That puts recruitment squarely in the mainstream of AI adoption. Companies from Chipotle to Unilever are reporting tangible results: faster hiring cycles, higher candidate completion rates, and measurable improvements in candidate experience.
So yes, AI is already here. The question is how far you let it run before candidate trust, fairness, or even basic accuracy take a hit.
The benefits recruiters are seeing from using AI
It’s safe to say the clearest win has been productivity. Recruiters who used to spend half a day buried in admin now get that time back. Joe said, “AI is brilliant for cutting out the boring bits, like summaries, notes, and first drafts. It’s not taking your job away, it’s just giving you more hours in the week to actually recruit.”
Coverage has improved too. “With AI you can run through thousands of CVs. That’s something no recruiter has time for,” Joe explained. “In theory, this could lead to some better outcomes for candidates for roles with 1000s of inbound applicants, now everyone has a chance of actually being screened for the role. However, there are some really key bias and compliance issues to consider for these tools, though.”
Outreach is starting to feel different as well. Writing tailored messages is one of those jobs that quietly eats your day. Joe told us, “If AI can give you a decent first draft, you can layer your own voice on top. It means you get quality outreach without sitting there for hours.”
| For more on using AI in recruitment, check out our guide to using Generative AI in recruitment, here. |
And sourcing, of course. “Boolean strings, market maps, competitor lists… It’s detailed work. AI won’t be perfect, but it can save a lot of the upfront manual work as well as building initial pipelines, allowing you to focus on the more niche talent pools,” Joe said.
The common thread is this: less time on basic admin and more time on the strategic work that actually moves a hire forward.
Risks and limitations to watch
For every task AI makes easier, there’s a catch. Bias is the one that always comes up first. Joe explains: “In the US, you’ve already seen lawsuits where AI systems filtered people out unfairly. And the important bit is the liability doesn’t just sit with the vendor. It’s the employer that’s responsible too.”
Data handling is another issue. Between GDPR and the incoming EU AI Act, recruiters can’t just experiment freely. “You can’t have everyone plugging in whatever tool they fancy,” Joe said. “There has to be governance. Who’s checking what data goes in, how it’s being stored, and whether it’s compliant? That’s the TA leader’s job to put in place, often working closely with internal compliance or security teams (where they exist).”
Candidate experience can also suffer if AI is left unchecked. “Bad AI outreach is almost worse than no outreach at all if it burns through potential candidates,” Joe added. “If your brand shows up in someone’s inbox sounding like a bot, they often switch off instantly.”
Similarly, there’s the risk of speeding up the wrong things. “If your hiring process is broken, AI just makes the broken parts faster,” Joe said. “You still need the basics in place, like role qualification, sourcing strategy, and calibration with hiring managers. Without that, you’re just automating the mess.”
So the warning isn’t “don’t use AI.” It’s: use it wisely, and make sure the guardrails are there before you scale.
How to integrate AI into your recruitment workflow (without losing the human touch)
Joe kept coming back to one point: AI should ideally slot into existing workflows. Here are four areas where it makes sense:
- Sourcing
AI is getting better at turning job descriptions into candidate lists, generating Boolean strings, or mapping out competitor talent. That’s a huge time-saver when you’re hiring in competitive markets. But the quality of results still varies. Recruiters need to treat AI outputs as a starting point, not the finished product. As Joe put it: “It’s great at surfacing candidates, not at telling you who’s genuinely a fit.” - Outreach
Drafting personalised outreach messages is one of the most time consuming jobs in recruitment. AI can help create the structure, pull in candidate-specific details, and give you a decent first draft. The trap is pushing AI-generated outreach out untouched. Often it’s better to treat it as a draft that saves you the blank-page time, and then refine with your own tone and context. In many cases, that extra polish is what gets a reply. - Screening
When you’re dealing with hundreds (sometimes thousands) of applicants, AI could make early screening more manageable. Pairing it with structured scorecards may speed things up and give you more consistent filtering. But you can’t take the recruiter out entirely. Motivation, soft skills, and cultural alignment are things AI isn’t ready to judge and getting those wrong can have big consequences. Joe was clear on this: “AI doesn’t know if someone’s motivation makes them right for the role, or whether they’ll fit the team.” - Internal adoption
AI works best when there are clear guardrails in place from the start, with room for recruiters to test and innovate inside them. Leaders can set the boundaries, such as compliance, data use, and candidate experience and then carve out space for teams to experiment safely. Some recruiting leaders we’ve spoken to even give their teams dedicated L&D time each week to play with AI tools, knowing that structured space is what turns curiosity into useful adoption. Joe described it as bottom-up meets top-down: “Let recruiters test tools, see what sticks. Then leaders can set the guardrails. That way, you get innovation without the chaos.”
Build vs. buy: choosing the right AI recruitment tools
When it comes to build vs buy, most teams end up somewhere in the middle. You don’t need to choose between building everything yourself or outsourcing the whole stack to vendors.
On the build side, recruiters are already using general AI or automation tools like ChatGPT, Zapier, or no-code automations to fix small pain points. Things like drafting Boolean strings, summarising kick-off calls, or generating outreach variations. These are lightweight experiments, fast to spin up, low cost, and flexible.
Buying is harder to avoid when it comes to data. Platforms like LinkedIn, or Juicebox have proprietary candidate pools you can’t replicate with a generic LLM.
The likely direction, as Joe sees it, is hybrid: “Most TA teams will probably end up mixing both. Use vendors for what only they can provide, then use your own AI solutions creatively in the gaps. That flexibility will matter, especially as the tools are evolving so quickly.”
What matters most is whether the tools actually make recruiters faster and candidates better served, without piling on complexity for the sake of it.
A day in the life of an AI-enabled recruiter
On a typical day, here are some of the key ways recruiters are already starting to use AI.
- Morning catch-up: Before the first call, your AI note-taker is already set up. It records, transcribes, and gives you a clean summary, which means no more scrambling through half-legible notes later.
- Outreach block: AI drafts the first draft of your InMails. It suggests personalisation points like a shared skill or a project someone worked on. You scan, edit, and send. What used to be ten messages an hour is now thirty.
- Screening window: Hundreds of applications come in. An AI filter runs them through structured criteria. You skim the shortlist, sense-check the edge cases, and spend your time on conversations rather than admin.
- End-of-day reporting: Instead of pulling data manually, an AI dashboard updates you on pipeline health, outreach response rates, and where bottlenecks are forming. You can walk into your sync with the hiring manager knowing the numbers cold.
None of this replaces the work it takes to be a recruiter, especially the conversational side of the role. You’re still the one who hears the candidate’s hesitation about relocation, or who convinces a passive engineer to take a call. What changes is the balance: less admin, more human work.
Making AI work for recruiters
AI helps recruiters get time back, widen their reach, and spend energy on the work that actually changes outcomes.
The trick is deciding where to let it run and where to step in. Use it for the admin, for the heavy lifting in screening, for the first draft of messages. But keep the parts that demand trust, judgment, and nuance firmly human.
Joe summed it up nicely: “AI should make recruiters more effective, not less relevant. If you use it well, it gives you space to do the things only a recruiter can do.”
If you’re tired of hiring constraints limiting growth, let’s talk. We’d genuinely love to help.
Want more insights on scaling your talent acquisition? Follow Scede on LinkedIn for strategies that actually work.