Enabling Is the Mode Most Often Missing
You can tell very quickly whether somebody is good at improving other people’s work. You cannot tell for a long time whether that person is any good at improving the people themselves. Organisations promote on the work output and then wonder why the team’s ability to get better seems stifled.
Improving the work is Editing. Improving the people is Coaching. In Describe the Work, Not the Title I described work as five kinds: Doing, Coaching, Editing, Navigating and Strategising. Cutting across them are three modes, set by the proportion of work you do and who that work is for. In Doing mode the outcomes are your own, and Doing is both a kind of work and a mode. In Enabling mode the outcomes are your team’s, and the work is mostly Coaching and Editing. In Framing mode you set the context everybody else works inside, and the work is mostly Navigating and Strategising. None of the three is a rung on a career ladder.
People who are good at Enabling are what I see in shortest supply, and their absence takes the longest to notice. Nothing breaks on the day Coaching stops. The work still ships, and the question of why nobody is getting better arrives much later, if it arrives at all.
AI tools are taking on more of the routine parts of the work, the building and the process repetition. That frees up capacity, and where it goes isn’t settled: it can turn into more output, or into the part of Enabling only a person can do. Capacity spent that way lets a good coach elevate a team faster than before. A team that improves slowly gets put down to people working hard; a team that visibly improves quickly makes everybody ask what changed. That doesn’t make Coaching measurable, but it does make good Coaching a lot harder to miss.
The shortage isn’t only about the wrong individuals being put into Enabling roles. It’s also the short-term, objectives-driven pressure inside organisations. Editing shows up immediately, and Coaching shows up over time, without a single defined moment to point to. An organisation running a quarterly or yearly review marks against set quantitative objectives first and has no way of gauging the five kinds of work, so it promotes on what it measures, not on the role somebody is moving into. It is why firms promote their best sellers and then watch those teams do worse, which Benson, Li and Shue found across 214 firms. Those firms weren’t choosing bad managers, they were choosing based on the only quantitative evidence in front of them. That is also why the freed-up capacity turns into more output by default. Spending it the other way has to be a decision somebody actually makes.
If a company hasn’t got people who are good at working in the Enabling mode, it has a problem it rarely diagnoses properly. Momentum never builds, people get frustrated and lose interest in the work, so those with options use them.
What AI Takes, and What It Doesn’t
There’s a reason these are modes and not just a description of work. Work can be handed over piece by piece, and AI is taking plenty of it. The quality of the work, and whether it was the right work at all, is still an individual’s responsibility when a person is delivering for others. So every mode splits in two: the part that can be handed over, and the part that can’t. AI is taking a great deal of the first. A lot of what gets written about AI at the moment reads as though the tools have intent of their own: the swarm enables, the system sets the direction, the agents work out what matters. They don’t. An AI can enable a swarm of sub-agents, and it can bring them into line with one another, but somewhere back along the way a person told the AI what to do.
In Doing mode, there are a lot of rote, repetitive tasks AI can do now, and it’s going to keep getting better at these. What stays with a person is starting the job in the first place, and the judgement to know what good actually looks like and what the work needs next. That matters most where the output isn’t binary, which is going to be most of the work worth doing in the future.
Framing is much the same. An AI can frame a plan really well, keep everybody current on it, and propose a sensible direction. But having the plan isn’t the job. AI also has a habit of drifting away from what was originally intended, so the judgement about whether it’s still the right plan, and the noticing when it isn’t, is what a person brings.
Enabling is the one I’d hold the line on, and I’d like the word back from the AI jargon while I’m at it. Enabling in an organisation is mostly Coaching and Editing. AI can do a lot of Editing well; the same can’t be said for Coaching. A coaching chatbot is sycophantic and it veers towards the middle, because that’s what it’s built to do. Making an AI your friendly coach isn’t the same as having somebody who has seen you in a hundred situations an AI never will, and who will tell you the thing you’d rather not hear.
The Newly Important Thing Keeps Changing
I joined ZURB, a product design firm in California, because I didn’t think I knew enough about designing good products. The product function I’d come up through ran on documents. Design was something else: making something meaningful enough that people would advocate for it.
While I was there I wrote that product leadership passes to whoever holds the newly important thing. At that point it was design judgement. Building features and functions had standardised to the point where they no longer separated one product from another; building a meaningful product did. That was one function, inside one industry, and I wasn’t the only one saying it.
It held for a while. Then the tools caught up, and making a product look and feel polished got a great deal easier. What that didn’t do was make the products any better at delivering meaning, which was never really about the tools. A feature can be finished. A piece of code can be finished. Design never is, and it isn’t binary either, which is why knowing what good design is stayed valuable long after the tools became better.
AI is the same situation, still just a toolset, not a final answer. The time you spend in the tools, working out how to get what you need from them, becomes part of your craft. It sits on top of whatever craft you already have rather than replacing it.
The building is getting commoditised again, and further than it was before. So it’s the Framing of a problem worth solving, the Enabling of the people around you, and the Doing that turns an output into an outcome somebody actually wanted, that are going to be the new important things. Even today, the hardest role to successfully fill in most of the organisations I’ve seen is the one that decides what’s worth building, not the one that builds it.
Hardest to fill and most often missing are two different scarcities. Framing is the hardest to hire for, because judgement about direction barely shows up in an interview. Enabling is the most often missing, because organisations rarely reward it or measure it effectively.
Whichever it turns out to be, the pattern keeps repeating. Design was simply the turn I happened to be standing in, and it has already passed on. AI tools will continue to eat into the work that can be handed over, but for the work itself a human still comes first on initial direction and final judgement. That’s the part that stays scarce, because it belongs to a particular situation and particular people.
How Open Are You to Working Differently?
The way this goes isn’t predetermined. AI arriving doesn’t mean somebody is coming for your job, even if that’s how it lands at first, and I understand why people may freeze at the thought of these changes. You can’t change progress though, so the question worth asking is a different one. How much are you held inside your current role, your responsibilities, your ways of working, and the professional identity you built on top of all of it? And how open are you to working differently: to learning, to helping the people around you learn, to bringing what you already know to something unfamiliar, to letting go of the part of the job that used to look like the important part, and being genuinely okay about that?
A great deal of Doing, as a kind of work, is now going to involve new learning. You have to work out how to do familiar work in an unfamiliar way, and there is no off-the-shelf AI package that fits your particular job or needs. So you build it yourself, and building it is Doing: hands on something new, learning it badly at first, gradually getting a feel for what good looks like. That’s a doer’s work, and for plenty of people at the minute it’s taking up more of their time than it used to.
Not everybody wants a route up through an organisation, and that’s completely fine. Some people are strong at the craft, happy enough to edit other people’s work, with no real interest in formally Coaching anybody through theirs. It’s the formal organisational responsibility they’re turning down, not the work. Plenty of people work in Doing mode for a whole career, and what they can achieve there, if they choose to, has changed and increased. And if you do want out of your current role, a new organisation or a one-person business running a swarm of agents is increasingly a genuine route rather than a consolation. None of those is a failure to progress.
These are still anxious times for plenty of people, and I don’t want to talk past that. The future isn’t written, though, and how this goes depends a great deal on what individuals decide to do with these tools, rather than on what the tools are going to do to them.
The next piece is about executing: how an operator in Enabling mode gets close enough to somebody’s work to see them accelerate, what changes when they do, and how that momentum compounds across the business.
Sources
- Alan Benson, Danielle Li and Kelly Shue, Promotions and the Peter Principle, The Quarterly Journal of Economics
https://academic.oup.com/qje/article/134/4/2085/5550760 - John Leenane, Design Just Ate My Software: How Designers Are Leading the Product Revolution, ZURB
https://zurb.com/blog/design-just-ate-my-software-how-designers
