All articles

Trends

25 thought-provoking ideas from Bill Boorman on recruitment and AI

Today’s recruiters are the last generation; in 6 years, there will be no more people, and job boards are suddenly back - but not as you might think. Bill Boorman is never short of controversial opinions and propositions. We have listed 25 recent ones.

placeholder 25 thought-provoking ideas from Bill Boorman on recruitment and AI

Hung Lee recently called him the ‘OG‘ of recruitment, and whoever you meet in the world of recruitment, you can never ignore Bill Boorman. The veteran of the profession, known among other things for the TRU gatherings, recently decided to share all his knowledge and vision via LinkedIn, at a dizzying pace, in so-called ‘Peace Out‘ posts. These are 25 of the most interesting insights on recruitment and A.I. that you could have gained from them so far.

#1. Senior candidates prefer the A.I. interviewer to junior candidates

The common assumption is that A.I. screening works well for high-volume roles and young candidates, but that experienced candidates and executives refuse it. Boorman sees precisely the opposite in data from agentic workflows. Senior candidates choose automated screening more often than a callback appointment with a recruiter. They do not believe that call will ever come, trust the neutrality of A.I. more than people, and want honest answers rather than a sales pitch, he believes.

In a real workflow, more than 90% choose A.I. when they have the choice.

Only once both sides are interested do they want to invest time and emotion in a conversation. Young people and junior candidates show slightly more resistance, not because they fear unfairness but because of the environmental impact of A.I. And strikingly, in surveys candidates overwhelmingly choose a human interview, but in a real workflow more than 90% choose A.I. when given the choice.

#2. A nurse is harder to find than a CTO

Agencies like to cite ‘hard-to-fill roles’ as their reason for existing, usually thinking of executives or other senior leaders. Boorman turns that around. According to his own ‘complexity to fill’ calculation, based on availability, scarcity and known talent pools, CEOs and CFOs are not difficult to find at all. Nurses, drivers, care workers and waiting staff are.

Executives are not difficult to find at all. Drivers are.

He recalls the coronavirus period, when ‘key workers’ were not wearing suits but driving lorries and keeping hospitals running. According to him, that is also where the best margins and greatest willingness to pay lie. Agentic workflows at Indeed and LinkedIn are therefore built from the frontline, not the boardroom. Indeed chief executive Hisayuki ‘Deko’ Idekoba spoke about lorry drivers in his pitch, not finance directors. So follow where the recruitment budgets are going.

#3. Job boards have become A.I.’s love baby

A statement Boorman says himself he never expected to make. Major financial institutions are approaching him to understand why job boards such as Indeed, LinkedIn and Seek are growing in a tight, stagnant market, whilst most companies have yet to turn their A.I. investments into growth. The answer: they no longer sell adverts but campaigns, linked to the outcome.

Indeed reportedly spends more on electricity than on tokens.

At Indeed, for example, average US revenue per vacancy rose by 35%, whilst the number of vacancies fell by around 4%. Crucial to this is proprietary data from both sides of the market, from Easy Apply to feedback from the ATS. Daily operations run on inexpensive in-house models; Indeed reportedly spends more on electricity than on tokens. Banks are now wondering whether this model works in every two-sided market.

#4. Agents mainly show how little humanity there was

The question of how to retain humanity when A.I. agents take over the work rests, according to Boorman, on a strange premise: that there is currently much humanity in recruitment. The black hole without feedback is the norm, and unconscious bias is widespread. He estimates that around 80% of tasks in the funnel can go to agents without noticeable loss, and that scale, speed and consistency actually deliver more human impact: every candidate receives feedback and a timeline.

‘Agents painfully reveal how much humanity was missing when we left everything to people.’

Teams that have already made the switch tell him that no time is freed up; new work fills the hours gained. They have delegated tasks but not decisions, and are now working on workforce planning, job design and internal mobility. His conclusion: agents will painfully expose just how much humanity was missing when we left everything to people.

#5. The real fraud problem is that nobody believes recruiters any more

When we think of fraud in recruitment, we quickly think of North Korean fake candidates or false applications (tech recruiters flag up to 45% as fake). The latter is now being reasonably addressed through authentication at the front end, according to Boorman. He even wonders whether excessive checking is not becoming as big a problem as the fraud itself.

His advice: restrict your channels until the scammers move on again.

The real damage lies elsewhere, he says: scammers imitate recruiters, using almost identical Gmail addresses and perfectly personalised A.I. messages. The result is that candidates see every unsolicited message as toxic, even if they know the brand and recruiter. Outreach through open channels is therefore becoming less effective. That is a blessing for paid, protected channels, because trust costs money. His advice: include it in your budget and restrict your channels until the scammers move on again.

#6. Let’s be honest: we do not care about the candidate experience

He says it is a deliberately provocative proposition. For decades, conferences have featured keynotes about improving the candidate experience, everyone applauds, and then… nothing happens. Crickets. If we had really cared, it would have been solved long ago, says Boorman, who says he was already a recruiter ‘before computers’.

‘If we had really cared about the candidate experience, it would have been solved long ago.’

He describes how every technological step led to new gates between recruiter and candidate. The funnel became a kind of Game of Thrones: the person ultimately hired was not the best, but the last survivor. Easy Apply was meant to solve that, but led to mass applications without reading the vacancy, and later to A.I. doing that for candidates. The result: nobody trusts anybody any more. Now that 93% of TA departments are in transformation, there is an opportunity. Take it, or stop talking about it, he stresses.

#7. A driving licence for every A.I. agent

What keeps Boorman awake: how will anyone still learn the profession? It is precisely the simple, clearly defined work through which junior staff learn that is automated first. He himself started in the 1980s by sifting through old CVs in a filing cabinet and was only later allowed to put forward candidates himself. Those mistakes were knocked out of him ‘not always gently’, sometimes literally, he says.

‘Nobody should rely on agents without first being skilled in the tasks that agent performs.’

If we soon hire only experienced recruiters for the work after the application, who will learn what happens under the bonnet of the agent? Without that knowledge, you can only accept what the agent says: you cannot challenge, train or improve it. His proposal: a licence for each agent. Nobody should rely on an agent without first demonstrably being skilled in the tasks that agent performs.

#8. People are quicker to stand up a machine than a person

The most important role of the ‘human in the loop’ is, according to Boorman, not compliance or bias-audits, but something much simpler: the social contract. An RPO director and the GM of JobGet separately told him the same thing. Candidates who arrange an interview through an agent conspicuously often fail to show up. JobGet therefore hired someone to call candidates and confirm the appointment, and expects to expand that team.

People feel less bound by promises made to a machine than those made to a person.

People feel much less bound by promises made to a machine than by promises made to a person. The first human in the loop therefore does not need great expertise, but above all a cheerful voice that briefly checks everything is all right. That would improve the figures enormously. His closing line: the social contract is not an algorithm.

#9. The CV is back, as a starting point rather than an end point

For years, the CV has been declared dead: hard to read in 20 seconds and hard to match on keywords. Yet job sites and talent networks are once again making CV upload mandatory as the first step, he observes. The reason? Contextual matching. The CV turns out to be the best starting point for a conversation: asking questions, qualifying, finding out what is missing. Exactly what a good recruiter used to do, and what agentic workflows now do.

Contextual matching brings the CV back to the centre of the process.

This shifts the principle from ‘Give us your CV so we can finish’ to ‘Give us your CV so we can begin’. Afterwards, the CV may never be looked at again, because the agent’s ranking and summary are sufficient. Until the hiring manager asks for it, because according to Boorman they still do not know how to conduct a conversation without a CV.

#10. Resignations say more than new jobs

After 44 years in the profession, Boorman concludes that he has always looked in the wrong place to understand the labour market. It is not the number of new jobs, but the number of people resigning themselves that is the real measure, in his view. If the so-called quit rate rises, the market is moving: replacement vacancies arise that are often urgent, and employers have to pay more for both retention and recruitment.

‘People have not suddenly fallen in love with their boss; they think: better the devil you know.’

New jobs, by contrast, can remain ‘sometime’ indefinitely. The current very low quit rate tells him that wages are stagnating and there is little confidence in the market. People have not suddenly fallen in love with their boss; they think: better the devil you know. Job-growth figures also conceal how many people combine multiple jobs to make ends meet. The conclusion for recruiters: your work will not become easier, only harder.

#11. Those arriving via ChatGPT are genuinely applying too

The owners of 4 job boards told Boorman the same story over 2 weeks: around 80% of visitors who arrive at their site via LLMs such as ChatGPT or Claude apply, and they complete that application too. It still accounts for only 5 to 10% of traffic, but that share is growing monthly. The explanation is intent: the visitor was in the middle of a conversation with the A.I., and the model concluded that this vacancy or employer was the answer.

Those who do not invest in GEO are missing the most motivated candidates, he says.

Most corporate sites are still impossible to find, however, Boorman observes, because they block A.I. crawlers. His practical tip: allow OAI-SearchBot in robots.txt, for example (that is different from GPTBot, which is for training), and write vacancies and content as answers to questions. Those who do not do so are missing the most motivated candidates, he says.

#12. The job advert now has 4 additional jobs

In the past, a job advert had just 1 task: stand out and get people to the apply button. Not any more. According to Indeed data, around 80% of hires do not happen because someone saw the vacancy, but because the system matched them to their profile and invited them. The advert is therefore first and foremost an entry ticket to the recommendation algorithm.

‘No wonder job adverts have become twice as long these days.’

It must then also contain enough structured data to match well with rich profiles, provide material for automatic outreach-messages, be discoverable through both contextual search and classic filters, and also appear in LLM answers. Only then comes its old task: persuading people to apply. No wonder job adverts have become twice as long these days, says Boorman. Recruiters, however, are trained in only 1 of those tasks, and therefore often fail at the rest.

#13. The hated auto-apply-agent is actually becoming valuable

An unpopular opinion, Boorman admits himself. Auto-apply-tools such as LazyApply are now seen as the main culprit behind overflowing pipelines and fake CVs. Rightly so, he believes, but that is because the first versions were sold to jobseekers for a few pounds a month. Adzuna co-founder Doug Monro pointed out to him: if the candidate pays, the incentive is to make as many applications as possible, and that is exactly what happened.

‘Agents are like teenagers at launch: wilful, boundary-pushing and full of mistakes.’

Agents are like teenagers at launch: wilful, boundary-pushing and full of mistakes. But they mature much faster than people, provided people guide them. The good variants now make candidates pay only with data and feedback, limit the number of applications per day, and require regular check-ins. He expects a world in which agents talk to agents, and would ideally like to see the option of auto-apply on every vacancy and careers site.

#14. In 6 years, there will be no more people

When Boorman first heard that the workforce was running out, he thought it was a warning for the next century. The speaker was talking about 2032. An ageing population that cannot or does not want to work on indefinitely, combined with ever fewer young entrants, is creating a structural supply problem, he says. This is already happening in the US, but even more strongly in Europe. The low mobility of people in work is already making it tangible.

‘Call me a conspiracy theorist, but bring data.’

His most striking conclusion: we desperately need the predicted and expected productivity gains from A.I., and should actually hope that A.I. takes over quite a few jobs, simply to avoid moving backwards. Raising the pension age will not solve it. He refers to Lightcast (the Fault Lines series) and Revelio Labs for the figures, and challenges critics: call me a conspiracy theorist, but bring data.

#15. Today’s recruiters are the last generation

He said it 2 years ago from a stage in Brno, and it went down badly at the time. Now, however, he hears more and more people reaching the same conclusion, although they phrase it more cautiously, wrapped in ‘new responsibilities’ and ‘broader tasks’. Agents and automation will take over around 80% of current recruitment work, and that is, according to Boorman, mainly the 80% recruiters did not enjoy anyway.

‘If only 20% of the role is still recruitment, is ‘recruiter’ still the right title?’

At the same time, responsibility and especially accountability are shifting towards internal mobility, career paths, retention and even internal communications. Recruiters now help determine which roles become human, which A.I., and which hybrid. If only 20% of the role is still recruitment, he wonders whether ‘recruiter’ is still the right title. For him, that is more than semantics: the new role is much closer to business goals than the old ‘logistical’ function. He sees it as good news and is already taking his hat off to it.

#16. Think in workflows, not categories

As an analyst and adviser, it was once his job to put tech start-ups in a box: which category, which competitors, which gap in core systems they filled. ATS, CRM and HRIS were the backbone, with separate tools clicking onto them like Lego bricks. That model no longer works, he says. A recent competitive analysis for an agentic tech company showed that it touched 6 categories without being the complete solution in any 1 of them.

ATS, CRM and HRIS were once the backbone, with separate tools clicking onto them like Lego bricks.

Customers are no longer looking for tools either; they are looking for workflows and measuring success by outcomes. Boorman has therefore thrown away all the old buying guides, quadrants and ecosystem maps. He now maps agents on the basis of all the tasks they perform, where that work begins and ends, and how far they get in the funnel before a handover is needed. Competitors are the parties that cover the same workflow. Agent against agent, not category against category.

#17. Why nobody buys your A.I.

Many A.I. start-ups in HR tech are in a painful position: a full pipeline, enthusiastic demos, but every deal stalls at the last moment. Boorman says he speaks to many desperate founders. Fear among large companies of security, compliance and mistakes is greater than the enthusiasm, partly because of upcoming legislation, bad examples in the news and the tendency to wait for someone else to do it first.

‘Buyers love an ongoing independent bias audit.’

The result: start-ups that run out of money and end in a fire sale. His advice: ensure you have a product that is easy to explain with demonstrable results, so serious pilots. Your partnerships team is more important than your sales team, because in 2026 everything revolves around distribution through the ecosystem. Be more online shop than high street, with a tile in the app store of the core systems. Build integrations quickly or use middleware, and adopt an ongoing independent bias audit. Buyers love it.

#18. DEI must start again, with inclusion for everyone

A friend of Boorman’s who worked on diversity lost his job in a public-sector restructuring, and he was not the only one. According to Boorman, around 40% of people with diversity or DEI in their job title no longer have that title, or no longer work for the same employer. Having diversity in your job title is now seen as toxic. As someone who has written thousands of words about it, he wonders why the post-coronavirus momentum did not create a sustainable foundation.

Having diversity in your job title is now seen as toxic.

His self-critical answer: we mostly told people what they should do, and listened too little to doubts. We underestimated that someone who is told they have an advantage does not readily give it up. We focused too much on specific groups. He argues for inclusion and fairness for everyone as the foundation, with concrete examples: accessible technology, processes, job design and marketing. Not sexy, but important.

#19. A.I. proves that the old recruitment rules are right

At 61, Boorman considers it good news above all: agentic hiring proves that the old-fashioned rules of the profession still stand. Those who break them will only get poor results on steroids with agents. The basis of every good hire remains a thorough understanding of the role and the candidate, and the overlap between them. Product people from start-ups to the biggest job boards tell him that the quality of the intake conversation between recruiter and hiring manager strongly determines matching and conversion.

‘The foundation remains a thorough understanding of both the role and the candidate, and the overlap between them.’

That conversation can be fully handled by an agent, but the real gain lies, in his view, in the human exchange, in which A.I. supplies live labour-market data such as salaries and talent availability so the recruiter can challenge more effectively. On the candidate side, agents handle screening at volume, and the human concludes the social contract. His summary: it is old-fashioned recruiting, but without the administration. Everything changed, and at the same time, nothing at all.

#20. A cap on applications: that is simply good recruitment

Recruitment groups are currently often debating whether you should close a vacancy once a maximum number of applications has been reached. The counterargument: perhaps number 101 is the best candidate. Boorman considers that comparable to endlessly hunting for the five-legged sheep. If you know your historical conversions, you also know how many applicants you need at minimum, and that is not rocket science, he says.

‘If you know your historical conversions, you also know how many applicants you need at minimum.’

With better A.I. screening and authentication at the front end, every candidate can at least be assessed and what comes through the front door converts much more efficiently. Where you once needed 100 applicants to speak to 10 and hire 1, you may now need only 5 to 10. That is not a reason to remove the cap, but precisely to close the vacancy once you have those 5. Good recruitment practice does not change with technology, only the noise around it.

#21. Every agent needs 5 requirements

Set an agent loose in your recruitment process and artificial intelligence quickly becomes natural stupidity, Boorman warns. Agents are different from the tools we grew up with: they are given a goal and find their own route to it, with access to email, databases and systems. They always answer confidently, persuasively but not always truthfully, and so quickly that the damage only becomes visible when it is too late. In recruitment, that can be costly and career-threatening, says Boorman.

‘Never trust a teenager completely.’

You delegate tasks, but responsibility remains with you. Together with VONQ CTO Bill Fischer, he came up with 5 requirements for putting agents to work properly: a clearly defined outcome, laws to operate within, limits to authority (and always asking when that limit is reached), truth above pleasing, and guidelines and standards. The golden rule on top of that: continually check and give feedback. What an agent does today, it may creatively do differently tomorrow. Never trust a teenager completely.

#22. Every technological revolution made recruitment worse

With every major technological change, from the internet to mobile and social media, the industry has, according to Boorman, missed the chance to make recruitment better. The more connected people became, the harder they proved to reach. The more data we gained, the better we became at collecting it and the worse at connecting. Now that the agentic revolution is coming, 2 challenges must set the direction, he says: trust and the availability of people.

‘Trust between recruiters and jobseekers is gone, and nothing can be repaired until it is restored.’

Trust between recruiters and jobseekers is gone, and nothing can be repaired until it is restored. According to him, this starts with sharing the gains fairly: productivity gains from A.I. should not automatically mean job losses. We need to agree on a basic decency in recruitment, even if it does not directly generate revenue or brand value. Fairness as a compass, not as a campaign slogan. And above all: do not repeat the old mistake by doing with agentic technology what we did digitally, what we once did on paper.

#23. Bring back old-fashioned training, precisely because nobody does the work any more

Boorman started in recruitment at sixteen, originally for a few weeks, and then remained active in it for 45 years. He was training director for 13 years at a national agency, in an era of smoky offices, signs with slogans such as Smile and Dial and managers who had never heard of an anti-bullying course. Yet even then, he saw the foundations that still apply today: a good intake, understanding what candidates really want, and preparing both candidate and hiring manager for the conversation.

Old school, new kool.

Two ears, one mouth. When he advised product companies such as TempBuddy, he simply handed out the whiteboards from his old office. Now that agents take over the heavy lifting, he fears those skills will disappear invisibly under the bonnet. He therefore calls on training companies to bring back that basic training. Not because the work is still done by hand, but because nobody will otherwise understand it any more. And who will then train and monitor the agents? Old school, new kool.

#24. Easy Apply became Lots Apply

A classic case of ‘be careful what you wish for’. To deal with candidates dropping out halfway through, applying has been reduced to 1 click in recent years. It worked, resulting in overflowing funnels and fraud. Archer by hackajob flagged 67% of applicants arriving through job boards and aggregators as suspicious. This has led to an entirely new pre-apply layer of screeners, light assessments and A.I. agents acting as gatekeepers for the ATS.

Can LinkedIn ever become the source of truth?

Boorman wonders whether that layer has by now completely undone the original benefit of Easy Apply: how often can you make someone prove they are real before they drop out? His expectation for the future is a portable, verified identity that follows the candidate everywhere, a kind of blue tick for applications. LinkedIn is trying to take that role, with verification becoming less and less voluntary. The question he asks: can LinkedIn ever become the source of truth?

#25. Do not simply ask people to talk to a machine

LinkedIn is replacing the familiar search bar with filters with the prompt ‘describe the job you want’, with examples such as ‘internships relating to climate issues’. An earlier attempt led to a flood of complaints about unexplained results, and Boorman often hears recruiters saying the same about the renewed search function in LinkedIn Recruiter.

‘Contextual search enables much better matches than a few keywords.’

He is enthusiastic about contextual search, which enables much better matches than a few keywords. But he believes platforms underestimate how big the transition is from entering information into a machine to having a conversation with it. Especially when looking for work: someone asked what they want will quickly think not of the honest answer, but of what the other person wants to hear. Anyone transforming websites, apps or job boards into contextual search must, in his view, consider how to help users through that change.

Bonus: There are only 3 metrics that really matter

Cost-per-hire, time to present and dozens of other KPIs are useful, but according to Boorman, everyone in recruitment really only needs to have 3 figures ready at all times: activity, efficiency and outcome. Activity is the raw volume at each stage, such as the number of applicants, screening conversations or candidate submissions. Efficiency is the conversion to the next stage. And outcome is what ultimately results. This immediately shows where things are getting stuck: too little activity, or activity of too low a quality to convert.

These figures are now hidden away in software, and: out of sight, out of mind…

They are also exactly the figures he always shares with tech teams to build and monitor agents, and in his view the best way to compare humans and agents fairly. His concern: these figures are now hidden away in software, and out of sight, out of mind. In the past, every recruiter reported them twice a day with pen and paper, meaning you could see problems coming before they became a real problem. So do not let your data be hidden in the tech, is his motto.

Read also

Read this article in full

Become a Werf& member for unlimited access to all articles, research and exclusive insights into the labour market.

Sign up now

No obligations Cancel anytime Secure payment

Related posts

All articles
placeholder How does political preference influence your selection? More than you might think…

Sustainability goes on your CV after Biden's victory, then comes off again after Trump's. American workers are adjusting their profiles in droves to match the political climate. And employers are rewarding that, even at the expense of quality. What does research say about political preference in recruitment?

Peter Boerman Peter Boerman Read article
placeholder How the practical training room is increasingly becoming the answer to staff shortages

Yesterday, another large practical training workshop opened in Tilburg, where the installers of tomorrow are being trained. Organisations are increasingly turning to their own practical training schools. Will they succeed in addressing the shortages?

Peter Boerman Peter Boerman Read article
placeholder How spy glasses can infiltrate your recruitment processes too

There is a great deal of debate worldwide about 'smart camera glasses'. This is also, or rather increasingly, true in recruitment. After all, who is really answering your questions? And what can you do about it?

Peter Boerman Peter Boerman Read article

Altijd op de hoogte van de ontwikkelingen op de arbeidsmarkt?

Sign up for our newsletter and receive the latest updates, news and invitations to our events.