What will your organisation become?
Same AI, very different organisations. A map for choosing on purpose instead of drifting, and where I pushed back on the AI that helped me draw it.
Picture two organisations that get the same AI. Both automate about ninety per cent of the same work.
One gives its staff a four-day week, and more time with the people it’s there for. The other cuts a third of its jobs.
Same technology, completely different organisations. The difference isn’t the AI. It’s what they chose to do with it, and who they chose it for.
Four analogies, borrowed from the high street
I started with shops, because everyone already understands them.
The artisan bakery does everything by hand and charges for the privilege. Its promise is the human craft itself. Lovely, if you can afford it; its tension is cost, scale and who gets left out.
The premium supermarket keeps the people you meet at the till and quietly automates everything behind them. The service you see is human; the jobs that go are the ones you don’t see. Its tension is that invisible displacement, and whether the human promise is still honest.
The small engineering firm keeps its core craft human and hands the rest to AI: HR, finance, scheduling. That works until AI starts moving the line on what counts as core.
The budget supermarket automates wherever it’s cheaper, because its customers want low prices more than a chat. Affordability and throughput go up; service and accountability are what get squeezed.
None of these is the “right” answer. They’re different propositions, with different costs, and different people paying them.
The newcomer
Then there’s the shape everyone talks about: the AI-native organisation. A founder, a handful of people, and agents doing most of the work.
It doesn’t really sit alongside the other four. They’re places an existing organisation can move towards. AI-native is where new organisations start. And that matters, because every shape now has an AI-native competitor: a budget supermarket up against a three-person firm running agents; a charity up against a new service bidding for the same grant with almost no staff.
Whatever you choose to become, you’ll be measured against something that started there.
Not a ladder: pulls
My first instinct was a ladder, from “all human” to “all AI”, with organisations climbing it at different speeds. The AI pushed back on that, and it was right. These aren’t stages, and nobody is obliged to climb. They’re more like attractors: shapes that organisations get pulled towards.
The pulls differ by sector. A business feels its margins and its competitors. A charity feels its funders and its mission. A public service feels its budget and its duty to everyone.
And one organisation can be several shapes at once. A law firm might be the engineering firm in its specialist advice and the budget supermarket in its routine contracts. A charity might be the premium supermarket in how it meets its supporters and the engineering firm in its specialist work.
Get Christian Graham’s stories in your inbox
Join Medium for free to get updates from this writer.
What worries me most is that most organisations won’t choose a shape at all. They’ll drift into one: a tool here, a subscription there, a pilot that quietly becomes the way things are done. And one day they’ll wake up as a budget supermarket that nobody decided to become.
A map to make the choice visible
Two questions do most of the work.
- Where does your value come from: efficiency, or the human touch?
- When AI saves money, who gets it: the owners, or the people? By “people” I mean staff, customers, or the people a charity serves.
Put those on two axes and the shapes spread out. The budget supermarket is efficiency and owners. The artisan bakery is human touch and people. The premium supermarket sells the human touch, but the gains go to its owners. And a co-op or employee-owned business can automate hard and still hand the gains back to its members.
The two organisations I started with both sit on the efficiency side. The difference is the second question: one handed the savings to its staff, the other to its owners. Same side of the map, different beneficiary.
There are other dimensions worth mapping if you want to go further: who actually makes decisions (staff, managers, communities or automated systems), who owns or governs the technology you depend on, and how much capability you really need. But those two questions are enough to start a useful argument.
An exercise. Ask your leadership team where they’d put your organisation on the map. Then ask your staff. If the two pins land in different places, that gap is the conversation you actually need to have.
The hard part isn’t the diagram
It’s people.
If you work somewhere, AI can look like a threat to your income, your identity and your say over your own work. That isn’t a misunderstanding to be fixed with better communications. Leaders and staff can have genuinely different interests, and reassurance without commitments won’t bridge that.
Both of the usual arguments are real, and both can be used to shut the conversation down. A business may feel it has no choice: if competitors automate and it doesn’t, there may be no jobs left to protect. A charity may feel every pound saved belongs to the cause.
Three things that seem to help
- Agree the principles before you run the pilots. How you’ll assess the impact on roles, how gains will be shared, what you’ll do if a job goes. Afterwards is too late.
- Give people real control over something that matters, like which tasks change first, or where the savings go. Participation should change outcomes, not decorate them.
- Measure more than speed. Workload, wellbeing, quality and the effect on the people you serve, alongside the efficiency numbers.
A workshop after the decision has already been made is consultation as theatre.
One question for your next team meeting
If you take one thing away, take this question:
If AI could do nearly everything we do today, what would we still choose to do with people? And what would we owe the people whose work changed?
Watch the video: What will your organisation become? (Organisations and AI, part 1)
Next: Who can switch your AI off? (part 2), on dependency, lock-in and the one number worth measuring.
Further reading
- Co-operatives UK: practical models for co-operative and employee ownership, the corner of the map where automation and shared gains can go together.
The organisations in this post are illustrative types, not real examples. Views are my own.
