Brokers modified every thing about hiring besides the result

Posted by techforbes June 19, 2026


Good Morning [%first_name |Dear Reader%],

You’re on a free plan.
Your subscription has expired.

Upgrade now to unlock premium newsletters, prime characteristic tales, unique podcasts, and extra.

The latest option to construct a brand new AI product is to say, “X is damaged, and we mounted it utilizing AI”. 

Properly, hiring was in all probability damaged, however AI didn’t change the result. 

As an alternative, it made it into an arms race the place candidates are bought AI instruments that assist them apply to firms and roles en masse in an automatic approach. On the opposite aspect, recruiters who’ve more and more gotten bored with the identical AI-generated solutions, resumes, and artefacts which were generated by AI have resorted to deploying AI methods to filter out candidates and get what they’re on the lookout for. And AI firms sit on either side of the transaction, billing tokens whereas hiring outcomes come again to precisely the place they have been.

Nearly precisely a yr in the past, I confronted this precise frustration. And I wrote about it in The Nutgraf in an version titled “AI sells swords at daybreak and shields at nightfall”.

I’ll return to the instance I initially shared—that of my Product Designer hiring. In an earlier world, I’d get 50 candidates, remove about 30 of them, and interview possibly 20. At present, I get 500 candidates, use AI to remove about 480 of them, and interview… 20 candidates. I’m again the place I began, besides each the candidates and I’ve paid AI firms for one thing neither of us signed up for.

The tragedy is that AI firms revenue from either side of the world they’ve created. On one hand, they’ll cost folks to create enormous volumes of low cost, low-quality content material, after which they’ll cost people to guard them from it. 

Web-net, we’ll find yourself in a world that’s not a lot totally different from earlier—besides that there’s a tax paid on either side. 

AI sells swords at daybreak and shields at nightfall, The Nutgraf

I’m fairly happy with the analogy I got here up with to explain this. Hiring must be extra like a “love marriage” with the candidate working with you as a live-in relationship. As an alternative, brokers have made it extra like an “organized marriage”, the place candidates have brokers who create their profiles, and brokers who consider different profiles, earlier than people even meet one another. 

Besides as an alternative of kinfolk and oldsters, the machines are doing the position, and getting paid for it. 

So in at present’s version of Two by Two, I made a decision to speak to 2 founders who’re on both aspect of this transaction—one in every of them is utilizing brokers and AI completely to rent, and the opposite is “promoting” AI to each candidates and corporations. 

To each of them, I requested the identical query: how does AI result in higher outcomes for each firms and the candidates? Aren’t we again to the place we started, besides with a token invoice? 

My first visitor is Aakash Dharmadhikari, co-founder and CPTO of Realfast.ai, a startup that’s an AI-native companies firm. Realfast caught my consideration as a result of they have been not too long ago hiring to scale up their crew and so they had a single distinctive requirement from all candidates—apply utilizing an AI agent. 

Image

In the meantime, on the candidate aspect, Saumil Tripathi, co-founder and CEO of Grapevine. Grapevine started as an nameless skilled community to debate salaries, layoffs, and office tradition. Then it swiftly grew previous a million customers. It has since pivoted arduous into AI hiring with two merchandise: TAL, an agent that scans job boards on the candidate’s behalf and Round1, an AI interview-simulation product. In some methods, Grapevine sits on either side of the AI within the hiring equation.

On this episode, each of those founders defend their approaches and argue that they’ve, in reality, solved hiring in their very own methods, with higher outcomes for firms and candidates. 

I’ll allow you to determine in the event that they persuade you about it.

1. The arms sellers don’t assume anybody wins this battle

Each founders constructed their firms on AI-hiring, and each spend the hour conceding the instruments are degrading the factor they’re paid to restore.

Saumil: “AI is definitely not fixing recruitment proper now. AI is within the strategy of destroying it. So, you understand, it should fully destroy recruitment earlier than somebody could make one thing that solves it.”

Aakash: “Should you’re going to develop quick and also you don’t need to make numerous errors, you must determine progressive methods of combating this arms race that you just’re speaking about—and the arms race half is totally true.”

2. When the take-home will be “one-shotted”, what’s left to check?

If a mannequin produces a cleaner resolution than your greatest engineer, the one factor value grading is the one half the mannequin can’t provide.

Aakash: “We’re not on the lookout for your coding skill. We’re truly on the lookout for your skill to be taught a totally new area. We’re your skill to drive an agent to provide sure artefacts.”

Saumil: “Past some extent, I have to say I nonetheless assume some p.c of persons are dishonest and we expect they’re very good and so they deserve a job anyway.”

3. The entire business filters out AI. One founder filters out the individuals who refuse it.

Recruiters are shopping for subscriptions to display AI-generated purposes, whereas these two are handing candidates extra firepower, not much less.

Aakash: “Fairly, when you don’t use AI, don’t apply.”

Saumil: “So all we did was we stated the businesses have a lot cash, a lot ammunition, allow us to give some to the candidates as properly. To not create extra noise, however to fall in love with a job and apply very deeply to that job.”

4. The indicators you’ve trusted for 20 years are those that died first

Resumes and canopy letters turned the best issues to pretend, so the true check moved to the dwell name and the agent log, i.e., the elements a candidate can’t pre-write.

Aakash: “E mail and resumes are now not unvarnished human work. That’s the drawback that we try to resolve.”

Saumil: “The candidates I gives you by way of Spherical 1 will likely be nearer to 90% or 95%. You’ll not hate the candidates I ship you—I’m not saying they’ll get employed, nevertheless it vastly will increase the sign aspect.”

5. Why the “simply come work with us” dream collapses above 4 years’ expertise 

Everybody romanticises proof-of-work hiring, however India’s discover durations and the two-way judgment of senior candidates squeeze it right into a slender band.

Saumil: “Someplace past three to 4 years of labor expertise hiring, this turns into a little bit bit impractical as a result of folks in India is not going to depart a job earlier than they get the following supply.”

Aakash: “So I really like the thought, I believe it’s doable inside a sure Goldilocks zone—it’s not going to occur past a sure scale, and it’ll not occur earlier than a sure scale.”

It’s an amazing episode. You possibly can take heed to it right here: “Agents changed everything about hiring except the outcome”

Para Divider

Do write to [email protected] or share your ideas within the feedback part. 

Regards,
Praveen Gopal Krishnan