Do AI recommendations convert to recruitment clients?

Yes they do, for some firms. But no one can yet give you the evidence you’re likely looking for.

Can anyone prove AI recommendations produce revenue?

Yes. The firms I spoke to last week have proven it to themselves.

Do firms winning AI recommendations want to prove it to the industry? Of course not. No firm worth their salt are going to be advertising results from this.

Three of the five firms I met last week had already taken client enquiries their clients put down to an AI recommendation. None of the three had done any AI visibility work. They knew because they asked at the point of enquiry.

One of them told me their client said it on the phone, unprompted:

“Copilot told me you are the best agency to speak to for this role.”

Why are some firms already getting this?

Recruitment firms that have done work on SEO have inherited decent GEO ground. If they have spent big bucks to rank on Google, they are naturally ranking on AI.

So why can’t anyone show you proof?

As this space is so infant, the measuring mechanisms simply don’t exist, and the only way to prove it works is seeing it for yourself.

I put the same question to ChatGPT, Copilot and Claude in August 2026. All three said the same thing without being asked.

ChatGPT: “I could not find a credible study establishing something like: ‘When an AI assistant recommends a recruitment agency, X% book a meeting.’”

Copilot: “Nobody can currently say: 100 AI mentions produce X meetings, X meetings produce Y placements. The market is too new.”

Here is the part that should make you sceptical of anyone selling this service, including me. You can currently put a number on AI impressions. You can put a number on a revenue figure. The move is putting them in the same paragraph and letting you join the dots.

Most of the conversion figures are based on a number of AI impressions, that led to website viewers, that filled in a form. Not a closed deal, or a new client.

When I talk about conversions, I talk about AI impressions becoming opportunities for your consultants to get in the room with them. Because that’s as far as my responsibility can stretch.

The biggest hole here is that no one can prove it, and that’s why the ground is wide open. Anyone that is proving it to themselves is milking it, not advertising it.

If you want to know who is telling you this and why, here is who I am and where I came from. If you want to see how I would look at your firm specifically, that is here.

Here is my data, and the holes in it

3.1% of my search impressions came from Google’s AI mode and overviews.

This points at the minimum amount of website visitors from AI, as it doesn’t track ChatGPT, Claude, Gemini or Perplexity. No consoles report all of that, and neither can anyone selling you GEO. Bing and GSC only launched beta AI analytics last month. This is a matter of time.

Right now, no one can give you an exact figure, and be wary of someone who is claiming they can. That being said, if you can verify what they can give as figures, then it’s the floor, not the ceiling.

I am poking holes in other businesses’ research, and in this service myself. That is the difference. I’ll show you my cards, not bullshit you.

How these numbers get bent

McKinsey published a figure saying a brand’s own website makes up only 5 to 10% of the sources AI search references.

“While SEO focuses on own-site content, in many cases, a brand’s own sites only comprise 5 to 10 percent of the sources that AI-search references.”
— McKinsey & Company, “New front door to the internet”, 16 October 2025

On the surface, it paints a picture that your success here is not attributed to your website’s presence, but instead your reputation with other domains. Interesting angle, from a business that sells that service.

Dive into the study and look at what it’s measuring. It is consumer buying decisions – cross-training shoes, credit cards, hotels, electronics, apparel. When you ask if a business is any good, the AI is of course going to search for reviews and endorsements to answer that question. What it doesn’t measure is when a hiring manager asks a specific question, about a specific role, in a specific place.

What it does tell us is that one of the largest consultancy firms in the world is seeing demand to answer how businesses get cited on AI. Industries are sitting up and taking notice. No longer looking at AI as a scary technology they don’t know how to handle, but as the next big customer acquisition channel it is becoming.

Then there’s LegalZoom, who published that AI traffic was up 250%, even though it remains 3% of their visitors.

Both data points true. There is a choice about which one you print:

Same company. Same quarter. Same data. Somebody chose.

That is exactly what is happening with every statistic on how you should use AI in recruitment, at a time when firm owners are feeling the pressure to keep up. This noise makes it impossible.

What the conversion numbers actually say

There are numbers out there, from different industries, and they are better than you are probably expecting by now.

Software companies, reporting their own signup sources:

Read that last one again. This shows either that way more of their traffic is AI than the 0.5% that is provable, or that AI visitors arrive with much more intent. My take is it’s a mix of both.

Software has been the first industry to tackle this, because their customers are in AI looking for solutions.

Another industry ahead of the game is legal. Interestingly, I found a lawyer at the start of this year through Claude, and it was spot on. It gave the perfect person to handle the case.

My advice?

There is no data or numbers that can keep you right during the shift we are experiencing right now. What can keep you right is your commercial acumen.

Everyone who is good in recruitment is like a shark with blood in water when it comes to commercial opportunities, and works close to the money. It’s very easy to become a busy fool in this space.

You are a leader in the recruitment industry. You have lasted long in this space because your instinct has kept you right. That will keep you right here, and that is what you need to listen to.

Navigating AI needs to be centred around making the right commercial calls, not analytical ones. Yes, data and research are important, but this is brand new technology realistically – your nose is the only thing you can truly rely on to keep you right. You don’t find success in recruitment from an Excel spreadsheet. You sniff out an increase in demand, under the radar, from speaking to clients and candidates, and you move on it before anyone has written it down.

You have to be clever, commercial and sharp to have success in this game, which I think is exactly what is needed to navigate the shifting world.

AI is already doing amazing things in recruitment for those that have joined these dots.

The real barriers to explore on whether this is for you

PSL businesses

A recommendation doesn’t mean you’re on the client’s PSL. In my market, relationship and delivery outweighed PSLs, and PSLs ultimately existed not as a guarantee of business, but as a guarantee of client favourable commercials in case you ended up working together.

It wouldn’t be a barrier for me, and honestly I don’t think for a PSL driven recruitment business either, as this exposure would include your PSL clients and potential PSL clients. But it definitely makes it a slower win.

Generalists

The generalist recruitment market is one the big players will be fighting over. Real success here is going to come from doubling down on what you already do really well, and becoming AI’s go-to firm for that role, industry and location. If you want to go wide, this might not have the same impact as it would for a firm going narrow.

How have early actors benefited from previous channels emerging?

In 2015 LinkedIn stopped recruiters from downloading PDFs of every company and contact on there for free.

The ones that had already done it owned that data for free, and had made inroads with them. The ones that hadn’t then had to pay for the opportunity to do the same, and wait until firms dropped the ball. They were left trying to win relationships other firms had formed. A task twice as hard, I’d say.

Now, no one can get past the second page of search results without a paid licence, and the recruitment tier, as you know, is pricey. This is because LinkedIn started to see the value it brought to the industry, and it pivoted to become a key tool for us.

Just like LinkedIn, AI is building dependency with everyone at the minute. At some point we are already starting to see the very start of monetisation around it. When that happens, that’s when this becomes a harder decision on whether the return is worth the spend.

ChatGPT, the largest platform, already carries the infancy of paid placement. Perplexity killed advertising, and Claude is ad-free by stated policy. Claude is now the largest share of enterprise AI spend, roughly 40% at the end of 2025, up from 24% a year earlier and 12% the year before that, and I suspect it’s grown more this year. This is the AI platform your clients’ businesses are buying, and it’s the one that currently has the biggest opportunity for conversions.

Right now, the tools don’t exist to prove it, so it’s cheap. As soon as the tools exist to prove it, everyone will pile in and the landscape becomes competitive. And 90% of the time there are shareholders behind this who will take notice and want in on the profits. Capitalism. It does the same every time.

For me, if I was still running a recruitment business, it would be a pretty easy decision.

Where I would start

Go and look at whether this is already happening to you. It costs nothing and it’s your data, not my anecdote.

In Google Search Console you can see whether your pages are turning up on AI surfaces. In Bing Webmaster Tools you can see how you’re being cited by Copilot – which matters, because Copilot sits inside the Microsoft account your clients are working in all day.

Then ask. Every enquiry, every time: how did you find us. Say AI out loud as one of the options. It’s the only attribution method that currently works, and in six months you’ll know more about this than any study will tell you.

One thing before you draw a conclusion from what you find. If you see some activity, that isn’t a clean bill of health. You’re being surfaced by a website built for how people searched two to five years ago. That was optimum. The ground moved.

Why I care about getting this right

In a world where people are actively trying to replace humans with AI, I want to look back on this period knowing I played a role in an industry that gave me my career, thriving, growing and everyone prospering.

That is why I’m doing this. To support the industry I’m passionate about, shift right.

To do that, you’re not going to get made up bullshit hot air data interpreted – to prove my point about what it’s actually doing. You’ll get encouraged to find and see the opportunity yourself. If you can, and you want to take it, then we could have a partnership here.

Common questions

Do AI recommendations actually produce recruitment clients?

Yes, for some firms. Three of the five UK recruitment firms I met in the week of 24 August 2026 had taken client enquiries their clients attributed to an AI recommendation. None had done any deliberate AI visibility work. They knew because they asked at the point of enquiry.

Can anyone prove the ROI?

No. The measuring mechanisms do not exist yet. Google Search Console only reports Google’s own AI surfaces, Bing Webmaster only covers Copilot, and neither reports what happens when someone reads an answer and picks up the phone. Anyone claiming a complete figure cannot have one.

Why are some recruitment firms already being recommended?

Because they have spent years on SEO. A firm that ranks well on Google is being read by the AI platforms too. That ground was bought, not won by luck.

Does this work if my clients run a PSL?

It is slower. A recommendation does not put you on the list. In my experience relationships and delivery outweigh PSLs anyway, and the exposure includes your PSL clients and your potential ones – but it makes it a slower win rather than a quick one.

What should I do first?

Check Google Search Console and Bing Webmaster Tools to see whether your pages are appearing on AI surfaces. Then ask every enquiry how they found you, and say AI out loud as one of the options. Both are free.

Darren Boyle

Founder, The AI Adoption Agency

Over a decade in recruitment before starting this. Now works with recruitment firms on how they get found and recommended by AI assistants. LinkedIn · About · Get in touch