Small and medium businesses are being left with no choice but to incorporate AI into their hiring processes now that the technology has enabled a volume of applications that humans alone can’t handle.
AI tools allow candidates to tailor resumés, generate cover letters and apply for dozens of roles at once, meaning many job ads attract hundreds of applications from largely unqualified candidates.
For a small business owner also managing recruitment, or a medium-sized business with a lean HR function, sifting through this number of applications is simply unworkable, says David Holland, Managing Director of Talent Solutions at Employment Hero.
He believes the technology that created the imbalance must also provide the solution. “For better or for worse, we absolutely have to have AI,” Holland says. His view is shared by HR leaders at businesses navigating the same pressures, including New Zealand facade systems manufacturer Thermosash and Gold Coast health food brand Macro Mike. In a webinar deep-dive, they cover the practical realities of AI in hiring: the time savings, the trade-offs, and the limits of what automation can replace.
Traditional Hiring Signals Are Disappearing
The first crack in the traditional hiring system appeared roughly two years ago for the Group General Manager of People & Capability at Thermosash, Craig Gillett. He recalls seeing the quality of applications change overnight.
“The standard of CVs and cover letters suddenly became incredibly impressive,” Gillett says. “CV after CV presented an amazing candidate who matched the job ad beautifully, whose cover letter linked the two together,” he said. “And I realised people were using AI to produce these high-value, high-volume, beautiful-looking CVs.”
A crucial step in the screening process was all but eliminated. “It became a lot harder to be able to quickly do that initial assessment and rely on the traditional indicators, like relevant experience or role fit capability,” he says.
Karli Newell, People and Culture Adviser at Macro Mike, says it’s become impossible to gauge which elements in a résumé or cover letter are real and which are AI spin. “There’s so much polish in applications now,” she says. “I think it’s taken a bit of the fun out of recruitment for me, actually.”
On Employment Hero’s platform, which facilitates roughly 600,000 applications per month across its client base, Holland notes a steady uptick in the proportion of candidates scoring 90 and above on automated matching tools, a direct result of applicants using large language models like ChatGPT, Copilot or Claude to optimise their submissions against job descriptions.
But he adds there’s an opposite cohort who no longer read the job description before applying, citing an industry anecdote that suggests as many as 40 per cent of candidates are using this spray and pray method.
“We’re losing the hiring signals,” he says. “In my view, there’s a technology war going on between the applicants who are rightly accessing technology to make their CVs unrejectable and those of us on the other side who are struggling to cope with the loss of signal and increase in volume.”
Leaders Are Reclaiming Hours Lost to Admin
But while AI has added to the small business owner’s capacity crisis in this regard, it is bringing about time dividends in others.
Gillett estimates that it used to take around 10 hours to move each candidate through to a second-stage interview, until AI intervened. “We can now make that happen in about 1.5 hours,” he says.
Part of the improvement comes from a prompt-based workflow his team built outside the HRIS. By uploading a candidate’s résumé, the job description and interview notes into a single prompt, the AI generates a candidate profile, an induction plan and a fit recommendation in less than 3 minutes. “Then the human brain comes in and you scan it, check it and make sure that it’s appropriately worded,” Gillett explains. He’s a strong advocate of refining prompts and has developed a library of prompts to speed up tasks throughout the employee life cycle.
Newell also points out that a distinction should be drawn between AI and automation. Automation, she argues, follows a fixed linear path, such as in payroll, where hours worked multiplied by an hourly rate produces a predictable output every time.
AI, by contrast, doesn’t follow a straight line. She uses chatbots as an illustration. “It’s not linear at all. That conversation could go left, right, and center through a maze based on the questions that I ask it and the direction it goes,” she explains. Unlike automation, the outputs aren’t always predictable in advance.
Trust and the Human Filter Still Matter
AI in hiring works best when it operates alongside humans and there is agreement on where lines should be drawn, says Holland. He advocates transparency, saying, for example, that if businesses use AI agents for customer support calls, they should introduce themselves as AI to begin.
He adds that AI offers consistency when dealing with a large volume of applications. “It’s actually able to reduce the amount of unconscious or subconscious bias,” he says. “It’s going to react to answers on a consistent basis. It’s going to give every candidate the same experience and opportunity to show the ‘real’ them.”
But Holland points out that human judgment will always be required later in the recruitment process. “I’m not going to suggest for a moment that we’re at the point that we want the AI to hire the person,” he says.
Newell highlights improvements in the candidate experience. She says AI can reduce the prevalence of ‘ghosting,’ a phenomenon that has frustrated more than half of job seekers. “There’s no excuse that a candidate shouldn’t hear from you because it should be done automatically.”
You can move people through stages and collaborate with hiring leads so much easier,” she says. Holland agrees, adding that a poor hiring experience can have reputational damage for an SME, even when it’s the result of overwhelm. “Leaving a candidate hanging for 40, 50, 60 days is the worst thing you can do.”
For SME owners who are keen to automate but unsure where to begin, Gillett offers a grounding principle: redesign your workflows before bringing in technology. “Don’t use AI to automate confusion,” he advises. “Build really clear processes on paper if you have to, and then use AI to make them faster, simpler, and more consistent.” Newell says a clear, top-down approach is needed to avoid fragmentation. “‘Are our teams getting so far down the track in their own individual ways of working that we don’t know what they’re doing and how they’re doing it?’” she suggests asking.
Gillett also shares a practical tip for getting better results from AI tools. “I’ve stopped typing just about everything,” he says. “I find the audio, even with the Kiwi accent, is really good.” His biggest trick: telling the AI what the output will be used for. “I’m going to use this information so that I can XYZ. And that transforms the quality of the output that you get back.”
His parting advice for small business owners still wary of AI is direct: “Get in and have a go. You can’t break it.”
