AI is creating the number 10 of editorial work

I’ve spent much of the past year trying to understand what AI actually changes about editorial work. I’m increasingly convinced one answer has very little to do with writing faster: it is that AI makes it possible to work between the lines.

Football has always had players whose value is difficult to describe using one position. The classic number 10 operated in the space between midfield and attack. Zinedine Zidane is probably my favourite example: not because he did everybody else’s job, but because he could read the shape around him, receive the ball in awkward spaces, connect phases of play and see passes other players had not yet seen.

They are valuable partly because they see the space between other people’s jobs.

I’m beginning to think something similar is happening to editorial work.

Most content organisations are still built around positions

Most organisations divide content into recognisable functions. Research. SEO. Social. Customer marketing. Editorial. Web. Analytics. Increasingly GEO and AI visibility.

Those divisions make sense. Specialists know things I don’t and can do things I can’t. I don’t think AI suddenly makes specialism irrelevant, and I am sceptical of the idea that one AI-enabled person can simply replace an entire content team.

But over the past year, some of the most useful work I’ve found myself doing hasn’t really belonged to any one of those boxes.

A GEO question leads me into customer evidence. A customer story reveals something that should change another piece of content. A video interview contains an answer nobody has extracted yet. Research supports an argument somewhere completely different. A product page exposes a knowledge gap rather than merely a copy problem. LinkedIn response tells me something about which ideas are finding an audience outside the one I already had.

None of those connections is especially extraordinary on its own.

The change is that I can now follow them.

Before AI, seeing the connection wasn’t enough

I probably would have noticed many of these relationships before AI. Editorial people have always been good at joining things up. Journalists move between sources, evidence, interviews and context all the time.

The practical problem was bandwidth.

Finding the old material takes time. Reading through ten customer stories takes time. Comparing a research report with a product page takes time. Going back through interview transcripts takes time. Analysing what changed between two versions of something takes time. Keeping enough context in your head to move between several projects at once takes a lot of time.

So even when I could see a potentially useful connection, there was often a strong incentive not to follow it very far. The cost of checking whether the instinct was right could become a project in itself.

AI has dramatically reduced that cost for me.

I can interrogate a body of material, compare sources, find patterns, preserve context, challenge an interpretation and move between projects much faster than I could before. That does not mean the machine decides which connection matters. In many cases it makes the judgement problem more important, because there are now far more possible connections available to inspect.

The valuable part is increasingly knowing which pass is worth attempting.

GEO made this particularly obvious

I’ve noticed this most clearly in the work I’ve been doing around GEO.

A GEO problem can initially look like a search or optimisation problem. There is a question an audience is asking and the organisation is not appearing as a useful answer. It would be easy to treat that as a job for the GEO person: analyse the gap, optimise a page, publish something new.

But the useful answer may already exist somewhere else.

It might be buried in a customer story. Somebody might have explained it in a video interview six months ago. Research might contain the proof but not in a form anybody can actually use. The product page might be trying to answer a question it was never designed to answer. Or the organisation might genuinely not know the answer yet, in which case the editorial task is not optimisation at all. Somebody needs to go and find out.

That makes GEO interesting to me because it exposes how fragmented organisational knowledge can be.

The job is not necessarily to invent a clever new GEO tactic. Sometimes it is to recognise that a GEO signal needs to be connected to research, customer evidence, product knowledge, editorial reporting or a completely different publishing surface.

The work is the connection.

That is also why I increasingly think of GEO as an editorial intelligence problem rather than only a distribution problem. The question is not simply whether a page is optimised for an answer engine. It is whether the organisation has a useful, evidenced answer in the first place, whether that answer can be found, and whether the learning survives long enough to improve the next piece of work.

This doesn’t make me everybody else’s specialist

There is an obvious danger in this argument. “Working between the lines” can become a flattering way of saying somebody knows a little about everything and not enough about anything.

That is not what I mean.

A number 10 does not become the goalkeeper because they can see the whole pitch. They do not replace the centre-back, winger or striker. Their value comes from understanding enough about the shape around them to connect different parts of the team.

I think the same distinction matters in editorial work.

I don’t need to become the best SEO specialist in the organisation to recognise when search data is pointing towards a missing editorial answer. I don’t need to own customer marketing to notice that a customer story contains evidence that should travel elsewhere. I don’t need to run research to understand when an argument needs external proof rather than another piece of copy. I don’t need to be the social lead to learn something useful from who a LinkedIn post reaches.

The skill is not universal expertise.

It is connective judgement.

That requires enough understanding of the neighbouring disciplines to know when something matters, enough organisational context to know where to look, and enough editorial judgement to decide what should happen next.

AI supplies leverage around that work. It does not supply the reason for doing it.

The new editorial role may sit between the boxes

For much of the past year I’ve been trying to work out where the kind of work I call Editorial Intelligence sits inside the existing content machine.

Is it strategy? Research? Editorial operations? Knowledge management? AI workflow design? Thought leadership? GEO?

Increasingly, I suspect the fact that it doesn’t fit cleanly into one of those categories may be part of the point.

AI is reducing some of the operational cost of crossing the boundaries between them. That changes what an individual editorial worker can realistically investigate and connect. It creates room for somebody who can move from a signal to the evidence behind it, from the evidence to an argument, from the argument to the right format, and from the response back into the organisation’s knowledge.

That feels different from the familiar idea of the “AI content generalist” who simply produces more things. The interesting role is not a one-person content department; it is closer to the number 10.

You don’t need them to replace everybody else on the pitch. You need them to see the space between the lines, connect what everyone else is doing and occasionally spot the pass nobody else saw.

For years we organised digital work around increasingly specialised positions. That was partly because specialist expertise matters and partly because moving across all those domains was expensive in time and attention.

AI changes the second part of that equation.

And if the cost of crossing functional boundaries keeps falling, some of the most valuable editorial people may be the ones with the judgement to know which boundaries are worth crossing.