A while ago I applied to be a trustee of a large national charity. I sent off my CV, it went through a sift, and a polite rejection came back. No conversation, no feedback. Completely normal, and fair enough: trustee roles are competitive, and someone has to make the call.
But it got me thinking. Having run my own businesses for most of my working life, I apply for things very rarely. And it showed me how hard it is to demonstrate value to a process designed around conventional careers.
I've led transformation and innovation work inside some big organisations. What I haven't got is a CV that climbs neatly from one recognisable title to the next. I've never been a Chief Transformation Officer. These days I write my own job titles, and they tend to change with my mood. Wonderful for me, less so for anyone trying to tick off a person specification at speed.
I don't think I'm unusual. I suspect lots of people with winding, unconventional careers drop out of sifts long before anyone notices what they might bring. Panels don't help either. They tend to agree fastest on candidates who look familiar to everyone in the room, so difference rarely gets argued out. It just disappears before the conversation even starts.
The queue is getting longer
Since I first started thinking about this, something has shifted, making the problem sharper.
Big organisations are shedding people as they restructure around AI. Many of those people are experienced and capable, and they're arriving on the job market at the same time. Meanwhile, the entry-level roles that used to absorb graduates are thinning out. The result is roles attracting hundreds of applications, and sometimes far more.
When volumes rise like that, sifts get blunter. They lean harder on keywords, titles and tidy progression, because that's what can be processed quickly. So the gap between a capable person and a job keeps widening. It's further now than it has ever been, both for the experienced professional whose career has taken a few interesting turns and for the graduate who hasn't had the chance to build one yet. Graduates tell me they send off hundreds of applications and hear nothing back, not even a no.
Which means organisations need a different kind of diversity. The people big organisations are shedding mostly reflect what's already inside them: similar careers, similar titles, similar ways of thinking. That makes them the easiest people to hire, because every sift recognises them instantly. So, the flood of applications doesn't widen the pool. It deepens the sameness, and pushes anyone with a less familiar route, or no route yet, further back in the queue. What organisations need now is range in how people got here and how they see the world, and that's exactly what most recruitment systems are least able to see.
Your recruitment process is a map
In Systemic Navigation, we start from a simple idea: people act on their maps of the world, and organisations do too. A recruitment process is an organisation's map of what good looks like, turned into a procedure. Job descriptions, competency frameworks, sift criteria and interview scores all encode it.
For a long time, that map fitted the territory well enough. When work was simple or complicated, you could define the problem, specify the expertise and measure performance against it. Fair, comparable processes were exactly the right tool.
The territory has moved on. More of what organisations face now is complex: messy, interdependent, political, with outcomes that only emerge over time. In that world, judgement matters more than execution, and the ability to hold several perspectives and stay curious under pressure becomes central.
Most recruitment hasn't kept up. It still assumes a person's contribution can be broken into neat CV lines: I delivered X, saved Y, led Z. That strips out almost everything interesting about how the work actually got done, the trade-offs, the trust, the learning. And as the old informal apprenticeship roles disappear, more people are building real judgement outside formal structures, through freelancing, side projects and lived experience. Their capability grows faster than their evidence for it.
Organisations, like people, tend to defend their maps until updating feels safer than defending. Recruitment is where you see that most clearly, because every sift is the map deciding what counts.
AI is holding up a mirror
AI is making all of this much harder to ignore.
I noticed it recently when hiring a web designer on Upwork. Every application was strong: thoughtful, well written, almost perfectly matched to my brief. Then I started talking to people, and the differences appeared. More often than not, the most impressive thing about an application had been the AI that helped write it.
I'm not knocking that. I use AI all the time, including on my own trustee application and to help shape this piece. When people are anxious to be noticed, they'll use whatever helps. But it does mean written applications are becoming a poor signal. Everyone sounds fluent and everyone fits the brief, so the thing you actually want to know about a person gets harder to see.
Most recruitment AI won't fix that, because it's built to slot into the existing process and keep decisions consistent and defensible. Define merit narrowly and AI will help everyone optimise for that narrow definition, only faster. Organisations say they want people who think systemically and handle ambiguity, then invest in the tools least able to spot either.
It could go another way. Imagine swapping the CV for a generative conversation, hosted by AI and designed to notice how someone thinks, what they pay attention to and how they handle uncertainty. Everyone curious about the organisation could have that first conversation. Curiosity would come first and filtering later, and the field would stay open long enough for something interesting to show up. There would still be bias, of course, but there's plenty in today's systems already. The difference is that you could write it down and examine it.
Welcome, as long as you fit
The same pattern shows up in diversity work. I'm fully behind commitments to equity and representation. The harder question is what happens once someone different arrives in a system that hasn't otherwise changed.
Too often, difference is welcome as long as it leaves things undisturbed. People are invited into the way things are done and rarely invited to change it. Representation increases while the range of ways of seeing remains much the same.
I've lived this. When I ran Kikass, a youth charity, organisations came to us precisely because we did things differently. That led to work with Virgin Unite, Richard Branson, and Number 10, as well as guerrilla stunts with brands like Levi's. When government brought us in for a major public health campaign, the reason was explicit: we could talk to young people in their own language.
Then the corridors of Whitehall got involved. Fuckitfairy and supershagland had been hugely successful for us, and they went down rather less well with civil servants. We were asked to swap our methods for safer ones, and the things that made the work effective were sanded away one by one. In the end we walked away. We'd been invited because we were different and then asked to stop being different.
So, what do we do with this?
I'll be honest: I don't have a tidy answer, and I'm wary of anyone who claims one.
Steven Bartlett has described a hiring test where candidates are given a scenario and a handful of sensible options. Most people pick one. A small minority ask why those are the only options. When AI can produce sensible, well-argued answers on demand, those question-challengers may be exactly who you need. They're also exactly who most sifts screen out.
Complex systems need enough people who think differently, in the right places at the right moments. Making recruitment fairer and more efficient still matters. The bigger challenge is building systems that can actually see the people they need.
If you hire, appoint or sit on panels, here are a few questions to take into your next round:
- What does your recruitment process assume good looks like, and when did anyone last check that against the territory you're actually facing?
- If everyone you hired this year could have been predicted from last year's organisation chart, what are you really building?
- When AI makes every application sound fluent, what can you still notice about how someone thinks?
- When someone different joins, what are they actually allowed to change?
- What would it take for updating your map to feel safer than defending it?
- What's one change to your next round, one degree away from how you do it now, that would let a different kind of person reach a conversation?
Some of what matters most can't be recruited for. It has to be recognised, and recognition starts with how we choose to see.
Neil Almond is Chief Systemic Navigator at 91Untold, where he helps organisations break the fixedness that keeps them stuck and create the conditions for movement. Join the conversation at www.91untold.com
More perspectives at 91untold.com/latest-thinking/
