Recruitment professionals share real-world applications of AI tools in hiring processes, highlighting the implementation, benefits, challenges, and advice.
Every recruiter navigates their own unique AI journey. Some focus on preventing candidates from misusing AI, while others actively explore its potential to enhance the hiring process. Meanwhile, some choose to observe from the sidelines, looking to learn and adapt as AI continues to evolve.
A LinkedIn survey found that while 57% of talent acquisition professionals are exploring or experimenting with Gen AI tools, 11% had already integrated them into their processes. ISE data shows that the arrival of ChatGPT has more than doubled the use of artificial intelligence in the early careers recruitment process. Yet views nor implementation are anything but uniform.
“Employers are at different places in their experiences and implementation of AI. AI is being used for things like drafting job descriptions, as well as driving process efficiency through keyword matching. But using AI to make recruitment decisions is a very different world, exposing employers to a whole range of legal considerations under UK GDPR and the EU AI Act,” explained Nicky Garcea, Co-Founder of Cappfinity.
The AI Act classifies HR-related uses of artificial intelligence, including recruitment processes, ‘high risk’. And concerns persist around the transparency of fairness and ethical implications.
“There is also the candidate perspective to consider,” says Robert Newry, Chief Explorer at Arctic Shores. “Students are regularly using generative AI tools and are increasingly proficient in their use, but many career sites are not clear on when and how such tools can be used in the application process.”
You’d be forgiven for taking a particularly cautious approach. Yet there are trailblazers. Companies like Amazon are leveraging machine learning to identify top candidates, assess genuine skills and abilities, and streamline processes.
Trailblazing
Amazon has considerable AI expertise and Lauren Gladwell, Recruitment Manager for EMEA Apprenticeships at Amazon, has been leading the approach in talent acquisition.
She started by building the team’s basic skills and experimenting, so they could become familiar with AI tools without compromising the integrity and effectiveness of the hiring process. This included linking up with their Global Talent Intelligence teams to make GenAI a continual focus of her team’s innovation programmes.
Several projects came out of these principles including the creation of an AI Risk Register and incorporating AI into their ATS to focus on skill-based selection processes. Following a successful UK pilot, there is an EMEA-wide proposal to scale the use of advanced assessment platforms, like Arctic Shores’ task-based assessment.
Read more about Amazon’s AI journey.
Experimentation
A cautiously optimistic outlook and desire to experiment are consistent features of the AI trailblazer.
In his first foray into AI, Jeff Lovejoy, a talent acquisition lead with a background in early careers hiring, saw ChatGPT revolutionise the screening process. It not only accelerated time-to-hire but also enhanced the quality of hires.
Jeff explained, “With talent shortages in particular markets, we expanded our job postings to a global audience, resulting in an overwhelming number of applications. To boost efficiency, our tech team proposed using ChatGPT for initial CV screening.”
Applications were processed through ChatGPT before entering the ATS. Criteria were based on the hiring manager's requirements. ChatGPT would analyse each CV and categorise candidates into four different buckets allowing recruiters to prioritise faster in a high-volume application process.
AI-generated summaries went into the ATS explaining the decisions, reducing the need for recruiters to read CVs.
“I became particularly excited when we implemented ChatGPT in our recruitment workflow,” explained Jeff. “This initial screening process streamlined our operations significantly, so we could focus on applications that truly required attention. Recruiters were able to manage their time accordingly across the different buckets, allowing for quicker responses to candidates and for the business.”
“It got to the point where it was so spot on that we became more and more confident in the decisions we were making. Data showed that candidates going into the different buckets aligned with what we would expect going through the pass-through rates of the different stages of our processes.
“We had a higher probability of success in our process. There were fewer candidates in our pipeline, but they were higher quality. We weren’t wasting time on people we wrongly assumed were a good fit. Ultimately, these tools have revolutionised how we approach recruitment, allowing us to focus more on quality over quantity and tailor our processes to hiring managers’ needs.”
Similar to Lauren, Jeff also had the advantage of being able to tap into in-house technical expertise. His tech team were keen to experiment, develop and were committed to driving it forward.
Not everybody has this level of support from within their organisation, so collaboration with specialist suppliers to leverage products as well as expertise can make using AI in hiring processes more feasible.
Leveraging third-party expertise
James Gordanifar, an expert in emerging talent acquisition - gained from global organisations such as EY, Clyde & Co, and WTW - used AI as part of the overall solution provided by a supplier.
He explained, “AI screened asynchronous video interview responses from over 30,000 candidates globally, where English was used as the primary language. It was part of a blended assessment, in which candidates completed cognitive games and answered questions that AI scored against the required core skills. Automation saved time prioritising top-tier candidates, and we had higher conversion rates at the final stages.”
Effectively using AI in recruitment reduces human error in interpreting CVs and by streamlining processes, early talent teams are freed up to spend more time making the decisions that count.
“Candidates were automatically placed into three tiers based on their performance, and this was where human interaction would come in,” said James. “Depending on the volume, recruiters would prioritise where to focus efforts and progress candidates to the final stages. To be clear, the AI was never intended to make a hiring decision; it was used as a tool to support recruiter efficiency as well as to indicate the likely success of the candidate in the role.”
The sentiment that it’s important to retain human decision-making authority is echoed by Jeff, who is currently working with an AI-featured ATS.
“Our tool doesn’t automate rejections; it highlights key skills and experiences to help us quickly identify promising candidates. Each CV is annotated with reasons for its assessment, allowing us to focus on areas needing clarification. For instance, if a candidate lacks particular skills, the system flags it, enabling targeted questions during interviews.
“AI should be creating efficiencies, not making formal hiring decisions. At the end of the day, this should still be a human making a decision.”
Bias concerns
While AI tools have greatly enhanced efficiency, they still have limitations. Three quarters of employers in the ISE Student Recruitment Survey said they worried about the potential for bias and preferred a more human-centric approach to recruitment.
Using AI in hiring could potentially discriminate against candidates warns Nicky. “There’s the potential for a detrimental impact on EDI, for example scoring video interviews around face recognition and emotion detection sensitivity has particular risk of impacts on underrepresented groups.”
However, there are ways to mitigate risk. James explained, “Risk has to be carefully managed and you’ll want to get into lots of data around adverse EDI impact. Any provider needs to be able to give you comfort that the tool will not introduce any form of bias. Our approach included ensuring that AI scores focused on the content of what the candidate said in their answer, not their facial expression, tone of voice etc. As a result, we benefited from solid diversity through minimal adverse impact.”
AI algorithms inherit biases from the data they are trained on. If historical data is biased, AI can reinforce or even magnify these biases, resulting in unfair hiring. To mitigate this, regular audits and diverse training data are crucial.
“If the AI was solely making decisions without us actually reviewing candidate profiles, then I'd probably have an issue that we were losing some potential diverse candidates. In our case, we're just asking, no matter what race or diversity or gender, we just need these skills. We continually monitored potential biases to ensure compliance with our D&I goals; ChatGPT helped identify top talent without disproportionately excluding minority or female candidates,” explained Jeff.
AI systems are designed to continuously acquire new knowledge and adapt to changing situations by constantly learning from new data and experiences, and this makes the future unpredictable. Robert cautioned, “As applications volumes jump and the AI enabled candidate gets more savvy, there will be a huge temptation for organisations to tackle AI with AI and given these models train on neurotypical models, there is a real danger under-represented groups will be once again unfairly treated.”
Advice for using AI in recruitment
For recruiters considering adopting AI in their processes, the consensus is to have a clear idea of the challenges you’re trying to solve, start slow, experiment and validate results before full implementation, maintaining human decision-making.
“For those considering adopting AI in recruitment, my advice is to start slow,” said Jeff. “Test the tools, review their outputs thoroughly, and refine the process before fully integrating them. AI should complement, not replace, human decision-making and don’t stop monitoring. Continuously review data to improve AI accuracy and maintain legal compliance.”
James reiterated the importance of legal support. “Risk means lots of heavy lifting to ensure compliance, especially on a global scale with a range of legal requirements. It is wise to consider whether your organisation is looking to globalise or is comfortable operating locally as this will have significant implications on feasibility of legal compliance.
“We had both internal and external council, and I would caution using a tool or an organisation that doesn't have deep expertise or credentials, as well as the capability to be part of the legal journey.”
What for the future?
There is optimistic caution around the use of AI in hiring. While there is potential to bring positive change and advancements, implementation requires careful consideration of potential risks and ethical implications, ensuring its benefits outweigh any potential harms.
“I think there is an obvious movement to explore what is possible and that is exciting, but just because you can doesn't mean you should. I think we have to get much better at questioning and being confident in the true value of what humans bring to a recruitment process and whatever we do, don't replace that, but use AI to enhance it,” said James.
There’s also the broader economic context to consider. Jeff commented, “Many professionals are hesitant, preferring to observe others before taking the plunge themselves. Some of the apprehension is concern over job security; if the implementation proves highly effective and efficient, it could potentially lead to reduced staffing needs. But, layoffs in talent acquisition often result from over-hiring or poor decisions. AI could help mitigate this by improving the quality of hires and reducing the frequency of costly hiring mistakes.”
ISE joint CEO, Stephen Isherwood considers the candidate experience and the consequence for employers, “I’d like to think that artificial intelligence will ultimately provide a better all-round experience for both the candidate and the employer, matching the right person to the right role with less wasted effort on both parts.
“But, as the technology develops in the immediate future, there is a danger that as students make an increasing number of applications, employer systems become increasingly impersonal to deal with the volume.”
Cappfinity has explained at length how employers might leverage AI in recruitment today and in the future. Nicky concluded, “Given EDI risks and ongoing legal controversy, we expect to see more use of AI to enhance recruitment efficiency in 2025, but with a careful delineation that it’s always the human who is making the decisions, even if those decisions are algorithmically automated through a human-designed process.”
The shift towards AI in recruitment is inevitable, but it comes with both challenges and opportunities. As the industry navigates these changes, balancing innovation with thoughtful implementation will be key to success. ISE is here to facilitate those discussions and share best practice.