By Matt Alder, host of the Recruiting Future podcast · Updated August 2026
Many TA functions are now deploying AI in some form. Far fewer are getting full value from it. After a decade of interviewing the people doing this work, including 276 episodes on AI, I think I know why. Recruiting was never deliberately designed. We inherited it, and it has barely changed in 200 years. AI is the first technology that could redesign it rather than just speed it up, but it amplifies whatever is already there. Organisations with well designed processes, clean data and a clearly defined problem are seeing real results. Organisations bolting AI onto an inherited process are just inheriting its failures faster. This page pulls together what works, what doesn’t, and what TA leaders should do next.
Three shifts come up again and again in my conversations.
The volume problem has flipped. Candidates are using AI to apply at scale, so recruiters are drowning in applications and resumes that all look the same. Meanwhile the best-fit candidates walk away from clunky application processes altogether, leaving TA teams overwhelmed and starved of the right applicants at the same time. I dug into this with Adecco and VONQ in Ep 711: How AI Agents Are Driving Recruiting Results. There’s a knock-on effect too: when thousands of near-identical applications arrive, filtering on actual predictive data is the only workable response, which is pushing proper assessment right to the top of the funnel.
Work itself is changing shape, not just recruiting. As Mark Stelzner put it in Ep 801: What Does AI-First Really Mean?, AI surfaces whatever was already underneath. Patchy data, conflicting content, processes nobody owns end to end. It all comes to the surface fast. Getting the data foundation right has become a discipline in its own right, which is why I explored it in Ep 809: The Data Foundation For AI In Hiring. Agentic AI is only as good as the data it can reach.
Discovery is moving inside AI tools. Candidates increasingly research employers and find jobs through LLMs rather than search engines and job boards. I looked at this in Ep 802: How LLMs Are Redefining Job Search. If your recruitment marketing is built purely for human search, it is quietly losing reach right now.
Recruiting has an innovation problem. It always has. The reason is not the one most people assume.
Underneath our processes, and underneath the systems that run them, sits a deeper layer: the mental models we all hold about how recruiting is supposed to work. Everyone in the workforce has been through a hiring process and absorbed a set of assumptions from it. There have to be resumes. There have to be interviews. Things happen in a certain order. We never address those assumptions directly, which is why wave after wave of technology has washed over recruiting without changing its underlying shape.
That was the argument I made in Ep 800: Will AI Break Recruiting?, and it changes what AI adoption means. We are at a fork in the road. AI could add value to recruiting at a scale we haven’t seen in 200 years, or it could be the point where the inherited process finally breaks under pressure it was never built for. Which way it goes has surprisingly little to do with the technology. It depends on whether TA leaders are willing to redesign the process and challenge the mental models underneath it.
The success stories on the show share a pattern, and it has very little to do with which tool was bought.
They lead with the problem, not the technology. When ACE Hardware and 7-Eleven transformed frontline hiring with AI (Ep 795: AI, Humans and Frontline Hiring), both teams started from a specific business problem, speed, and decided deliberately where humans stayed firmly in the process.
They fix the process before they accelerate it. Toni de Graaf of the Recruiting Excellence Foundation (Ep 808: The Starting Point For TA Innovation) is blunt about this. If you can’t explain exactly how your selection process finds the best ten candidates out of a thousand, you shouldn’t be putting AI anywhere near it. His foundation has benchmarked TA maturity globally across 21 components, and the weakest areas everywhere are the job kickoff meeting and the hiring decision, which are exactly the judgement-heavy moments where TA adds the most value.
They redesign rather than automate. One of the strongest case studies in the archive is Adecco’s agent-led process (Ep 711). An AI screener that reads every word of every application, asks follow-up questions, scores transparently, and surfaces transferable skills the old process would never have found. Around 90% of candidates opted in to AI screening, because a human alternative was always on the table. That isn’t the inherited process running faster. It’s a different process.
They decide what humans do with the time saved. Automating candidate review might free up half a sourcer’s week. If nobody designs what that time gets redirected into, the strategic work everyone talks about never happens. This came through strongly in Ep 808 and echoes across dozens of implementation conversations. Capacity only becomes value when someone architects it.
Candidates are sceptical about AI in hiring, and the research backs them up. In ThriveMap’s study, 49% of candidates thought AI in recruiting was unfair, against only 22% who thought it was fair. Whenever AI screening makes the news, the story tends to lead with the negative.
That scepticism deserves an honest answer, and the honest answer is uncomfortable. The status quo we’re comparing AI against is not some golden age of human attention. Millions of applications are never read by anyone at all. So the real question is whether a transparent AI screen that gives every applicant a response, and a route to a human, is worse than silence.
Honesty is the way through this. Candidates want to know when and how AI is being used on them, and transparency in AI interviewing and assessment is quickly becoming a reputational issue rather than a compliance footnote. That’s the subject of Ep 807: Trust, Transparency And The AI Interview. The employers getting this right, like the frontline stories in Ep 795 and the Adecco case in Ep 711, are building processes where AI visibly improves the candidate’s experience. Candidates choose it, rather than having it done to them.
Guest after guest draws the same line, whether they’re a CHRO, a founder or a frontline TA leader. AI takes on the rote work: scheduling, screening, admin, first-pass review. Judgement, relationships, influence and the high-trust moments stay with humans. The kickoff conversation with a hiring manager. The hiring decision. Executive alignment. Change management. The point of automating the funnel is better human connections, not no human connections. If anything, AI raises the premium on these skills, because they’re now the part of the role that can’t be commoditised. Kelly A Monahan put a useful discipline on it in Ep 777: Why AI Needs To Drive Value Not Efficiency: decide what you put a fence around. The decisions that stay human no matter how good the technology gets.
Distilling the practical advice from across the archive, five priorities determine whether AI delivers real value.
Underneath all five sits the same capability: seeing what’s coming early enough to act calmly. That’s why Foresight comes first in the FITT framework I introduced in Ep 700: Are You Fit For The Future?.
I hold predictions on the show to a discipline: separate what’s true now from what’s still vendor positioning. What’s true now is that AI is already changing how candidates apply, how employers screen, and how both sides find each other. What’s coming, on a longer timeline than the hype suggests: agentic workflows running more of the funnel, quite possibly with candidate agents and employer agents talking directly to each other, assessment moving to the top of the process, and one provocative thesis from Ep 808, that hiring volume rises as organisations shift towards specialist and gig hiring for AI-orchestrated work. The TA teams that thrive will be the ones who learned to redesign and iterate. Not the ones who bought the most tools.
Browse all 276 episodes tagged AI.
Recruiting Future is an independent podcast for senior talent acquisition and HR professionals, hosted by Matt Alder since 2015. It publishes two episodes a week plus a monthly Round Up, and has published 882 episodes exploring AI, recruiting technology, recruitment marketing, employer branding, skills-based hiring, assessment and the future of work. Full details are on the about page.
Matt Alder is a talent acquisition futurist with over 25 years of experience across talent acquisition and talent strategy. He works with enterprise TA and HR leaders on AI readiness, has delivered keynotes in 18 countries, and is the co-author of Exceptional Talent and Digital Talent, both published by Kogan Page with Mervyn Dinnen.
Recruiting Future is an independent production. Sponsors do not set the show’s agenda or its editorial positions, and the show does not rank or recommend vendors.
