The Organisation is the Product

We all know this.

Deep down, we all know this.

We’ve spent the best part of twenty years trying to make Product, Engineering and Design more agile.

We changed the teams.
We changed the ceremonies.
We changed the job titles.
Squads. Tribes. Chapters. Missions. OKRs. Value streams.

And yet decisions still take too long. Priorities still multiply. Teams still spend extraordinary amounts of time navigating the organisation rather than solving the problem.

We changed the teams. We didn’t really change the organisation around them.

So we spent twenty years sending people on one-day Product Owner courses. We learned SAFe. We introduced new frameworks, processes, ceremonies and terminology.

And somewhere along the way, something rather ironic happened.

We became slaves to the process that was supposed to make us agile.

And now we’re about to do it again

This time with AI.
We’re buying copilots. Deploying agents. Automating tasks. Building AI strategies. Creating AI functions.
All perfectly sensible.

But there’s a danger that we’re doing exactly what we did with Agile.

Or worse, someone will come up with the AI equivalent of the Spotify model and we’ll all copy that instead.
The Spot-AI-fy model, presumably.

New boxes. New terminology. New job titles. New diagrams.
Same organisation underneath.

We’re putting a new way of working into an organisation that was designed for the old one.
And then wondering why the transformation doesn’t happen.

Maybe that’s because we’ve spent years optimising individual parts of the organisation, without really redesigning the system they operate in.

We optimised the parts, not the whole

Product people are taught to start with outcomes.
Understand the problem. Remove unnecessary complexity. Make ownership clear. Test assumptions. Learn. Iterate.

Then we walk into organisations that seem to operate by completely different rules.

A customer problem crosses five departments.
A decision requires three committees.
A team owns an outcome but not the systems, people or budget required to deliver it.

When ownership isn’t clear, we add a role.
When functions don’t coordinate, we add a committee.
When decisions aren’t being made, we add governance.
When something still doesn’t work, we add another team.

If this were a product, would we ship it?
Probably not.

Yet somehow we’ve accepted that this is simply how large organisations work.

Organisations accumulate debt too

We understand technical debt.
We understand product debt.
Organisations accumulate debt in exactly the same way.

A temporary role becomes permanent.
A committee created after an incident is still meeting five years later.
A process designed for an exception becomes the standard process.
A new function appears because two existing functions can’t agree who owns something.

None of these decisions is necessarily stupid. Most probably made perfect sense at the time.
That’s how debt works.

Eventually you end up with an organisation nobody would ever have designed deliberately.
It just happened.
Layer by layer. Reorganisation by reorganisation. Workaround by workaround.

And now we’re about to add AI to it.

So design the work, not the jobs

Instead of asking where we can put AI, perhaps we should start somewhere different.

• What outcome are we trying to create?
• What work creates it?
• What intelligence does that work require?
• What context is needed?
• What decisions have to be made?
• How should people, agents and technology coordinate?
• Where does accountability sit?

Only then should we decide what the organisation looks like.

Some work will remain human.
Some will be automated.
Some will be performed by agents.
Much of it will probably involve all three.

The interesting question isn’t how many jobs AI replaces.
It’s how we redesign the system through which work gets done.

The Organisation is the Product

For twenty years, we’ve become increasingly sophisticated at designing digital products.
We observe them. Measure them. Find friction. Remove complexity. Experiment. Learn. Improve.

Perhaps it’s time we applied the same thinking to the organisations that create them.

Because AI isn’t simply another technology we’re introducing into the enterprise.
It changes what work can be done, who — or what — can do it, how decisions get made, how people and technology coordinate, and where accountability sits.

Every organisation will increasingly have access to the same models, agents and technology.
The advantage won’t simply come from having AI.
It will come from designing an organisation that knows what to do with it.

The Organisation is the Product.
Perhaps it’s time we started designing it like one.