What a recruiter is worth now

The public conversation about AI and recruitment has been stuck on the wrong question for two years. Whether the job survives was never seriously in doubt to anybody who has sat on a desk, and we have argued that at length in will AI replace recruiters. The interesting question is the commercial one nobody wants to open, which is what a recruiter is now worth, and it has a specific answer that is good news for the person doing the job.

A recruiter has never been paid for finding people. They have been paid for judgement applied to a situation: whether this person is right for this business, whether this client is telling you the whole story, whether this candidate is about to make a decision they will regret. The finding, the typing, the chasing and the sorting were the cost of getting that judgement into contact with a situation. They were overhead, and overhead is what has just been removed.

So the value of a recruiter did not change because they got better. It changed because the number of situations their judgement can reach in a week went up by an order of magnitude, and nothing else about the supply of that judgement changed at all.

What is a recruiter worth now

More than they were, because the scarce thing in the process is now the only thing left in it. This is not encouragement, it is how prices work when one input becomes abundant and another does not.

Finding candidates used to be genuinely difficult and it was most of what an agency charged for. It has become close to free, and anything close to free stops being worth a fee. What has not become free is a person who can read a hiring manager, tell a client something they do not want to hear, judge whether someone will survive a demanding team, and be accountable for the recommendation. The supply of that has not increased by a single unit, and the demand for it is now the whole of the demand.

The mistake is to read that as the job shrinking to a smaller core. What actually happened is the opposite. The core stayed the same size and the thing around it fell away, so the same person now spends their week almost entirely inside the part they are good at. A consultant who was exceptional with people used to get Thursday afternoon for that and Monday to Wednesday for data entry. That ratio inverted, and the exceptional part now gets the week.

What exactly changed in what a client is buying

The fee is the same fee and it is now buying something different, which is worth being precise about because it changes how you sell.

What the fee used to cover	What it covers now
Access to candidates nobody else could find	Access to candidates who will actually take the role
The labour of assembling a shortlist	The judgement behind a shortlist of four
Keeping the process moving by remembering	Keeping the process moving by knowing when to push
Reading CVs against a spec	Reading people against a team
Coverage of a market by volume	Knowledge of a market by relationship
Being the agency with the biggest database	Being the person who tells the client the truth
Effort the client could see	Outcomes the client can count

Read the right-hand column and notice that every row is harder to fake and harder to copy than the row beside it. That is the commercial shift. The left-hand column could be competed away by anyone willing to hire more juniors. The right-hand column cannot be competed away at all, and the agencies pricing against it are having a better year than the ones still selling the left.

The limits are real and they are on the same side of the line. What a machine cannot do on a desk is the subject of what AI cannot do on a recruitment desk, and every item on that list is in the right-hand column here. The permanent parts of the job and the valuable parts of the job turn out to be the same parts, which is a fortunate arrangement for recruiters and an awkward one for anybody selling their replacement.

How does this show up in billings per head

As the only number that settles the argument, and most agencies are not yet tracking it properly. Activity is up everywhere this year. Billings per consultant is the figure that tells you whether any of that activity reached the outcome, and it is the one an owner should be able to state without looking it up.

Take total fee income for a period, divide by the number of fee-earning heads, and hold it against the same figure two years ago. That is the whole calculation. If it has not moved while your activity and your tooling spend have both risen substantially, the hours you reclaimed went into the front of the pipe rather than into anything that produces a placement. That failure has a mechanism and we set it out in the pipe doesn't care where you widened it.

Two supporting numbers make it honest. Live roles per consultant tells you whether the ceiling actually moved, because a desk still holding twelve jobs has not changed however much software it runs. Fill rate tells you whether the extra roles are real, because a hundred roles at a poor fill rate is a worse business than thirty at a good one, and the transition to a portfolio can quietly trade one for the other if nobody is watching.

I would add one more, which is the number I would look at first if I were buying an agency. Fee income per consultant, split by whether the consultant was in the top or bottom half of the team last year. If the gap widened, the tooling amplified the people who were already good, which is what it does. If the gap narrowed, the good people are being held back by something, and that something is usually a process that still requires them to do admin the software could have taken.

Who captures the gain

This is the question the industry is avoiding, and it will be answered in the next couple of years whether anybody plans for it or not. When one consultant's output rises several times over, there are four parties who could keep the difference, and only one of them has to be persuaded.

The client, through lower fees. Some of it will go here and some of it should, because a market where the cost of finding people collapsed will pass part of that on. The important thing is which part. A fee reduction justified by cheaper sourcing is defensible. A fee reduction justified by the idea that recruitment as a whole got easier is not, because the part the client is actually buying did not get easier or cheaper at all.

The agency, through margin. The default outcome, and reasonable to a point, since the agency is carrying the tooling cost and the risk of the transition. It becomes unreasonable when it is the whole of the answer for more than a year, because the consultant will do the arithmetic.

The consultant, through pay. The one that most agencies have not touched, and the one I would move first. A consultant producing several times what they produced two years ago, on a commission structure written for the old output, will work out their position quickly. They have more options than they have ever had, because the same collapse in overhead that let them run a hundred roles also lowered the cost of running their own small agency to almost nothing.

Nobody, through waste. The most common outcome so far, where the reclaimed hours went into more sourcing and the gain never materialised anywhere. It is not a distribution question at that point, because there is nothing to distribute.

What happens if the gain is not shared

Your best people leave, and they leave to compete with you rather than to join someone else. This is the part of the transition I would be thinking hardest about if I ran an agency rather than a software company.

The barrier to a good consultant starting their own desk used to be operational rather than commercial. They could always recruit. What they could not do alone was the collecting, typing, chasing and sorting that came with every job, which is why a strong biller needed an agency around them. That barrier is most of the way gone, and a single consultant with good tooling can now carry a portfolio that would have required a small team five years ago.

So the retention argument has to be made on something else. Clients, brand, colleagues, the work you win that an individual cannot, and a share of the gain that reflects what they actually produce now. Agencies that hold their best people through this will be agencies that repriced the job. The ones that keep a structure written for a ten-job desk while asking for fifty-job output are going to find out how portable their talent has become.

There is a version of this that goes well for everybody, and it is not complicated. A consultant producing more earns more, the agency earns more in absolute terms on a larger base, and the client gets a shortlist of four people who will actually take the job instead of a longlist of forty who might. The distribution is a choice, and delaying the choice is itself a choice.

What should a recruiter do about their own worth

Four things, and they are all within reach of an individual rather than requiring permission.

Know your own number. Fee income you personally generated, and live roles you personally held, this year against two years ago. Very few consultants can state both, and the one who can is in a different conversation at review time from the one who cannot.

Let the admin go, completely. Partial adoption is the worst position, because you carry the cost of the new system and keep the old work. Every hour you spend typing something the system already knows is an hour of the most valuable thing you own spent on the cheapest thing in the process. What that desk looks like once the transition is complete is in a 10x recruiter runs a hundred jobs.

Put the hours where they compound. Client relationships and the candidates already in your process. Both build something that is yours next year as well as this one, which is more than can be said for a larger longlist.

Get better at the part that cannot be copied. Reading people, knowing a market properly, telling clients uncomfortable truths, holding a candidate through an offer. These were always the valuable skills and they were always the ones that got whatever time was left. They now get the week, and the gap between a consultant who has worked on them and one who has not is about to become very visible in the billings.

The honest limit on all of this

Your client still interviews two or three people a week, and no amount of capability upstream changes that number. It is set by a hiring manager's diary and it has not moved in twenty years. Everything in this piece about a recruiter being worth more depends on the reclaimed hours landing on that constraint rather than in front of it.

Which is why the worth of a recruiter went up rather than the worth of a process going up. The bottleneck in hiring is a human decision, made by a person who wants somebody they trust standing next to it and answerable for it. Every tool that clears work away from that moment makes the person standing there more important, not less, and there is no version of this technology that changes who has to make the call.

That is the whole argument, and it is the reason I am more confident about this industry than about almost any other. AI will never replace recruiters. What it has done is take the ten job orders that filled a week and make them a hundred, because the admin stopped being yours, and the judgement that was always the point now reaches every one of them.

We build our AI and the agents on that belief rather than against it: the machine takes the admin, it reports how sure it is, and anything it is not confident about comes back to a recruiter to decide. Recruitly is the best recruiting CRM in the world, and we built it that way because the recruiter is the part of this that is actually worth something, and it is getting more valuable every year.

---

Gowri is the CEO of Recruitly.