In this piece
I sat on an APSCo webinar panel recently and ended up debating one of the other panellists. The view put forward was that recruitment businesses scale on average billers, not top performers.
It made me ask whether the market rewards average billers any more.
That question sits at the centre of what’s happening in recruitment. This isn’t another cyclical downturn where we batten down the hatches and wait for the bounce. The UK has seen 36 consecutive months of declining vacancies. The industry has shed 30% of its workforce in just over a year. AI is already changing how the work gets done, while many agencies are still running the old playbook.
I run a job board, recruit in the education and training sector, and advise agencies on automation and process. So I’m not theorising about what might work. I see what does work, up close, every day.
Efficiency first
If you’re spending 60% of your gross margin on people in a recruitment business, you won’t be in business in two years.
That’s not hyperbole. JP Morgan grew productivity 10-20% last year while cutting headcount. The link between headcount and productivity, which has held since the Industrial Revolution, is breaking. The banks figured this out first, but it’s coming for everyone.
One of my clients, a boutique agency, is replacing a full-time contract biller with AI-assisted workflows. Productivity hasn’t dropped. It’s improved.
The maths is brutal. Your cost base needs to shift from 60% people to 40% people plus digital workers over the next 12-24 months. That’s survival, not a nice-to-have.
Most agencies are going to the same vendors to solve this, buying the same tools their competitors buy. At best, if you integrate it well, you’re treading water. At worst, you’ve just added overhead while telling yourself you’re innovating.
Personal brand is overtaking company brand
There was a moment in the webinar where I mentioned a solo recruiter billing £800k. Just him. A small brand behind him. No corporate infrastructure.
That’s not an outlier anymore. I think personal brand has already outstripped company brand for everyone except the Korn Ferry tier, where customers are buying the status signal. The “nobody gets fired for using IBM” insurance policy.
Below that? Clients follow people, not logos. The best recruiters I know could move tomorrow and take their business with them. This is already happening. What matters is whether agencies are set up to keep these people, or are just getting what they can out of them before they walk.
I agree with my fellow panellist that you can’t scale a business on prima donnas that don’t follow the rules. But why aren’t we building processes around top performers, instead of forcing them into workflows designed for the average biller?
If someone’s billing multiples of their peers, give them an assistant. Build them custom tools. Treat it like football agents managing talent, not factory workers on a production line.
The agencies that figure this out will keep their best people. The rest will watch them leave and wonder why.
Still, knowledge can’t live in one person’s head if you want to drive enterprise value. That doesn’t mean letting top billers do whatever they want. It means wrapping them in support systems that spread their best practice across the business while keeping them free to perform. They’ll be happy, and so will you.
To retain that knowledge, L&D must be an apprenticeship model. Grey hair coaching the next generation, knowledge transfer as a core function. Not hoping one in ten junior hires figures it out by making 100 cold calls a day, while the rest burn out.
Your data is your moat (not your LinkedIn licence)
Everyone on the webinar kept circling back to data. One panellist made the point that if you don’t know your baseline metrics, you can’t measure whether AI is helping.
Consider that everyone has access to the same AI. ChatGPT, Claude, whatever. So why doesn’t that level the playing field?
Because AI runs on your data, and your data is different.
Your CRM holds relationship context, placement history, which clients respond to which candidates and in what timeframes. That’s proprietary intelligence. LinkedIn Recruiter gives you public profiles. Useful, but it’s the same public profiles your competitors see.
The agencies winning right now are tracking:
- Time to shortlist (ideally under 72 hours, and one panellist cited data showing a 2x success rate in the first three days)
- Fillability scores per role type
- Cost per CV that actually converts
- Client response time as a leading indicator
None of these are vanity metrics. They show you which roles to work, which to park, which clients move fast and which waste your time. That’s the moat.
SMEs have an advantage here, weirdly. They can move faster, build more targeted systems. They’re not trying to serve everyone. Some of the most profitable recruitment happens with smaller clients who move in 5 days versus enterprise PSLs that take 8 weeks to process.
Build vs buy: why I’m contrarian on rectech
During the webinar, we touched on building. A fellow panellist stated that he’d never seen a small agency successfully custom build.
I get his point. He’s coming from enterprise transformation roles where he’s seen millions wasted on failed builds. I’ve never seen it done either, but I think it’s far more achievable now than it was.
No-code platforms, API integrations and AI-assisted development make small builds possible in a way they weren’t five years ago. You’re not rebuilding Bullhorn. You’re solving a specific problem your competitors aren’t solving.
One client wanted to automate contractor availability checks. Manual calls every morning, a huge admin drain. We built a simple workflow that texts contractors, collects the replies and routes urgent gaps to consultants. It cost a few hundred quid and freed up hours a day. I’ve written up more workflows like it in AI for recruitment firms.
That’s not on the shelf at Rec Expo. And because it’s custom, it’s defensible.
Shadow IT, where consultants sign up for random tools behind IT’s back, isn’t a problem to stamp out. It’s a signal. Your best people are working around your systems because the systems don’t serve them. Listen to that.
Look for the desire lines in your business.
Integration matters. Everything needs to talk to everything else. But integration doesn’t stop you building small, targeted tools on top of your core stack.
The real question is whether a build lifts your margin or traps you in overhead.
The practitioner’s advantage
Being hands-on still matters. When I’m advising agencies, I’m not guessing what might work. I’m testing it in my own businesses first.
That creates a different conversation. I can say “here’s what broke when we tried it” or “here’s the workaround we found” because I’ve actually done it.
So here’s a test for you. Track where your consultants spend their time for one week. If 60%+ is admin, sourcing, or staring at LinkedIn, you’ve got a problem. Because AI can do all of that. Not as well as your best person, but better than your average person, and at a fraction of the cost.
The path forward is moving your people onto the work only humans can do: reading the room, building trust, delivering bad news well, and understanding the emotional drag of a career move.
That’s the work AI won’t touch for years. But you can’t get there if your best communicators are still chained to a desk running Boolean searches.
What happens next
The webinar ran long because we kept circling back to the same tension: everyone knows change is necessary, but knowing and doing are different.
Some agencies are in denial. They think the market will bounce back and this will all blow over. It won’t.
Others are paralysed. Too many options, don’t know where to start, drowning in tech fatigue.
The ones who’ll win are treating this like a forcing function. Not waiting for permission or perfect information. Pick one efficiency lever (candidate response times, role triage, data enrichment) and obsess over it until it works better.
Then move to the next one.
Start somewhere. This week, not next quarter. Because while you’re deciding, your best people are deciding too. They’re wondering whether to stay or whether they can do this better with someone else or on their own.
And the maths on that decision is getting easier for them every day.
If you want to start with one workflow, such as availability checks or role triage, book a 30-minute call. I won’t promise to agree with you, but I’ll give you an honest view.
